NEUROPHYSIOLOGY
NEUROPHYSIOLOGY
NEUROPHYSIOLOGY
Essentials of Clinical
Neurophysiology
Third edition
Karl E. Misulis, MD,PhD & Thomas C. Head, MD
Neurology, Semmes-Murphey Clinic, Vanderbilt University,
& University of Tennessee, Memphis
Some of the membrane proteins are channels which allow certain ions to pass. Other
proteins are pumps which actively push ions through the membranes against a
concentration gradient. The sodium-potassium pump pushes sodium out of the cell while
pushing potassium into the cell. This creates a chemical gradient for sodium to go into the
cell and potassium out of the cell. Without the channel proteins, virtually none of the ions
would move in or out of the cell,
however, the channel proteins
allow for controlled movement of
the ions.
Membrane potentials
Differential permeability of cell membranes to ions produces a membrane potential. The
sodium-potassium pump does not pump the cations exactly 1-for-1, so there is a slight
membrane potential created by this discrepancy. However, the largest component of the
membrane potential is related to efflux of potassium from the cytoplasm to the
extracellular space.
At rest, the permeability of the neuronal membrane to K+ is greater than to any other ion.
Therefore, K+ diffuses out of the cell down its chemical gradient. Anionic proteins cannot
accompany K+ through the channel, so they are left behind. This results in the interior of
the cell becoming negative compared to the exterior.
The potential being created across of the cell membrane opposes further flow of
potassium down the chemical gradient. This electrical gradient opposes the flow of K+
down the chemical gradient. When the electrical gradient is equal in strength to the
chemical gradient, flow of potassium ceases. The system is in equilibrium and this is
termed the equilibrium potential. For K+, the equilibrium potential is -75 mV.
Equilibrium potential: The equilibrium potential is the potential at which the electrical
gradient is sufficient to exactly counter-balance the chemical gradient.
The equilibrium potential for each ion can be calculated from the Nernst equation:
[ K ]i
E K = −58 log
[ K ]o
[ Na ]i
ENa = −58 log
[ Na ]o
[Cl ]o
ECl = −58 log
[Cl ]i
Where EK, ENa, and ECl are the equilibrium potentials for sodium, potassium, and
chloride, respectively. the bracketed letters are the concentrations of the ions with the
following i and o representing the concentrations inside and outside of the cell,
respectively. The concentrations for chloride are reversed, with the outside concentration
over the inside concentration because of the negative change of chloride. There are other
ions, but these are the most important for the resting membrane potential.
is dependent on the permeability of the membrane to that ion. This is described by the
Goldman constant field equation:
GK [ K ]i
V = −58 log
GK [ K ]o
[ K ]i
V = −58 log
[ K ]o
Figure 1-2: Conductance changes with an action potential.
Sequence ofmembrane potential and ionic conductance changes
during an action potential, Potassium conductance predominates
This is essentially the same as
during rest, Sodium conductance is markedly increased during the equilibrium potential for
the action potential. K+. Therefore, the resting
potential approaches -75 mV
but is not exactly that value
because of a small contribution from Cl- and Na+.
Generator potentials
Generator potentials are produced by changes in conductance, so that the Goldman
constant field equation results in a potential which differs from the equilibrium potential
of potassium.
• Transmitter stimulation
• Electrotonic conduction from surrounding membrane
• Mechanical deformation
• Leaky membranes
Neurotransmitters are released from the presynaptic terminal and bind to receptors on the
post-synaptic membrane. Binding results in an increased conductance to sodium and
calcium which then depolarizes the membrane. This is the generator potential.
Some neurotransmitters
Figure 1-3: Electrotonic conduction. activate channels for
A: Lipid bilayer membrane traversed by a sodium channel, Opening of potassium and chloride.
this channel results in depolarization which is maximal at that site, but Influx of chloride into the
which is also present downstream from the site of channel opening.
cell plus efflux of
B: Graph of the depolarization as a function of distance from the site.
The decay is exponential.
potassium out of the cell
makes the membrane
potential more close to
the normal resting membrane potential. This acts to inhibit the effects of generator
potentials, since an increase in sodium conductance is no longer able to produce the same
degree of depolarization. While this is not a generator potential, the physiology of
membrane channel conductance makes discussion here appropriate.
Electrotonic conduction
Depolarization occurs at a site on the membrane with a certain geography. The
depolarization affects adjacent membrane, however. If one records the membrane
potential at sites progressively farther from a site of depolarization, the magnitude of the
depolarization at that site will be less.
ECN 3/e - 7
Electrotonic conduction not only works within a cell, but can also work between cells if
there is a tight junction which allows for electrotonic conduction from cell to cell. This is
termed an ephapse. In the next cell, the depolarization may be sufficient to generate an
action potential even if there was no action potential in the first cell.
Action potentials
Action potentials are transient depolarizations which are sufficient to make the membrane
potential positive. Not all neurons have action potentials, some conduct electronically,
with depolarization of one region of the cell having an effect on depolarizing adjacent
membrane without an action potential. All peripheral nerves and major relay nerves of
the brain have action potentials. The phases of the action potential are:
• Generator potential
• Regenerative depolarization
• Peak potential
• Repolarization
• After-hyperpolarization
The generator potential was discussed in the previous section. The regenerative
depolarization is due to activation of voltage-gated sodium channels. These channels are
opened by the membrane potential reaching a certain voltage level which differs for each
channel. If the potential threshold is crossed, the channel briefly opens. The increased
sodium conductance produces further depolarization, and so on. Eventually, all of the
sodium channels which can open have opened, so the membrane potential reaches a peak
potential which is near the equilibrium potential for sodium, about +45 mV.
Repolarization is
accomplished by closing of
the sodium channels and
opening of potassium
channels. The closing of the
sodium channels is because
they are time-dependent,
that is, they only have a
fixed duration of opening.
When the sodium
conductance decreases, the
Goldman constant field
equation reverts to
simplification such that the
Figure 1-4: Action potential propogation: Myelinated vs membrane potential is close
unmyelinated axon. to the equilibrium potential
A: Unmyelinated axon with propagation of the action potential for potassium. This effect is
down adjacent axonal membrane.
augmented by opening of
B: Myelinated axon with propagation of the action potential from
node to node, thereby conducting faster than in A. potassium channels. In fact,
ECN 3/e - 8
Myelinated axons
conduct more quickly
than unmyelinated
axons because of the
myelin sheath. The
sheath serves several
purposes, but the main
purpose is to decrease
the ionic leak of the
axon. As shown in
Figure 1-4, the
depolarization is
conducted further down
the axon when there is a
myelin sheath than
Figure 1-5: Synaptic transmission. when the axon is
A: Synaptic transmission with release of transmitter from the presynaptic unmyelinated. Since it
terminal and binding of the transmitter to receptors on the postsynaptic takes time for the action
membrane.
potential to be
B: Electrotonic conduction of depolarization from synaptic transmission
down the postsynaptic membrane. generated, the
C: Excitatory post-synaptic potential (EPSP) produced by synaptic myelination results in
transmission. The EPSP on the left did not reach threshold for action faster conduction down
potential generation, but the one on the right did. the axon.
Neurotransmitters
Transmitter release
ECN 3/e - 9
Not all neurons develop action potentials. If the axon has an action potential, a fixed
amount of transmitter is released each time. If the neuron does not develop an action
potential, then the amount of transmitter released is graded, dependent on the degree of
depolarization of the terminal.
Receptor binding
Transmitter diffuses across the narrow synaptic cleft and binds to the receptor for the
transmitter on the post-synaptic membrane. Binding to the receptors produces the effect
of the transmitter. Typically, the receptor binding produces opening of membrane ionic
channels.
Post-synaptic potentials
Receptor activation produces excitatory post-synaptic potentials (EPSP) or inhibitory
post-synaptic potentials (IPSP). EPSPs are due to increased conductance to sodium,
calcium, or both. EPSPs sum to allow the adjacent membrane to reach threshold.
IPSPs are due to increased conductance to potassium and chloride. This clamps the
membrane potential in the negative range, so that the membrane cannot reach threshold
for action potential generation or even for ion influx.
Muscle physiology
Neuromuscular
transmission
Neuromuscular transmission is
essentially similar to synaptic
transmission as described above.
Action potential propagation
down the motoneuron axon
produces release of acetylcholine
from the nerve terminal. The
acetylcholine crosses the small
gap between nerve and muscle
Figure 1-6: Muscle contraction. and binds to the acetylcholine
A: Muscle fiber in the relaxed state. receptor. This opens channels in
B: Muscle fiber contracted. Crosslinking between actin and the muscle which allow for influx
myosin filaments shortens the muscle. of sodium and calcium. These
ECN 3/e - 10
channels are more permissive than many other channels, which have specific ions which
are allowed. The movement of so many ions results in loss of the negative resting
membrane potential, which in turn triggers an action potential in the adjacent muscle
fiber membrane.
Muscle contraction
Calcium and ATP are used to facilitate repeated cross-linking and releasing of the actin
and myosin filaments. Calcium is essential for the cyclical binding and release of the
filaments. Re-uptake of the calcium into the sarcoplasmic reticulum produces relaxation
of the muscle.
Most electrophysiologic recordings are of either nerve bundles or nerve connections. All
recordings are extracellular, so intracellular recordings will not be considered.
Nerve bundles
Nerve bundles in the PNS or CNS conduct action potentials. For many studies, there are
synchronous action potentials in a nerve bundle, such as in nerve conduction velocity
measurements or evoked potentials where the nerve was electrically stimulated in one
location and recordings made from another location. Extracellular recordings, particularly
made through the skin, are unable to detect a single action potential, so a synchronous
volley is essential to detection.
Nerve connections
ECN 3/e - 11
Nerves have synapses mainly in the cerebral cortex, cerebellar cortex, and nuclei. In the
cerebral cortex, the synapses are oriented throughout the layers. Major cortical efferent
neurons are oriented vertically with dendritic arborizations such that excitatory or
inhibitory activity can produce an electric field within the cortex which has a vertical axis
and a radial distribution.
Most of the electric activity recorded during EEG is synaptic activity whereas the signal
recorded during measurement of nerve conduction is nerve conduction, i.e. action
potentials.
Scalp EEG recordings cannot “see” a single neuron, so activity has to be synchronous
over many neurons before a wave is produced. Even then, it is attenuated through passage
through skull and scalp. Seizure activity is discussed in Chapter 3, but typically differs
from most non-epileptic activity by the synchrony of the electric potentials.
Near-field potentials
Near-field electrodes are recorded
by electrodes close to the
membrane, whereas far-field
potentials are recorded at a distance.
To illustrate, consider a peripheral
nerve conducting a volley of action
potentials. Refer to Figures 1-7
through 1–9 for this discussion. The
near-field potential can be recorded
by a unipolar or bipolar recording
arrangement. For a unipolar
recording, the active electrode (G1
Figure 1-7: Unipolar near-field potential. for Grid 1 from old tube
A: The wave of depolarization has not reached the recording
electrodes.
terminology) is directly overlying
B: The depolarization is maximal at the active electrode, the he nerve, and the reference (G2) is
inward current producing a negative extracellular potential. on the same limb but not overlying
C: The depolarization has passed the active electrode, so the the nerve. As the wave of
potential is returning toward baseline. depolarization passes under G1, the
D: The depolarization has passed so the recorded potential is
at baseline. The reference electrode is electrically inactive,
inward current causes a prominent
since it is on a distant region. negative wave. As the wave passes,
the potential returns to baseline.
Far-field potentials
Bipolar recording is similar to a unipolar recording except that G2 is over a distal
segment of the nerve. The initial negative component is as previously described. As the
depolarization passes under G2, however, a positive phase occurs because the amplifier is
measuring the difference in potential between G1 and G2. Depolarization at G1 makes
G1 negative relative to G2. Depolarization at G2 makes G2 negative relative to G1, but
this means that G1 is positive relative to G2. Therefore, the recording is biphasic. The
far-field potential is recorded with G1 at a distance from the current generator and G2 at a
greater distance, typically in a different tissue compartment. The far-field potential is a
stationary wave that is due to the moving front of depolarization in the axons. Since the
far-field potential is recorded at a distance, it is not governed by exact electrode position.
If the nerve volley comes close to the active or reference electrode, a component of near-
field potential can be seen.
Most of the potentials recorded in clinical neurophysiology are far-field potentials. The
notable exception is the compound action potentials of nerve conduction studies, which
are largely near-field potentials.
ECN 3/e - 13
Conductors are materials which have partially-filled orbitals which can accept electrons
and have loosely-held electrons which can be donated to nearby atoms. Conductors may
be metale, some synthetic materials with conducting atoms, and ionic fluids such as
biologic tissues.
Non-conductors do not
have partially-filled
electron orbitals or
loosely-held electrons.
Therefore, even strong
electric or magnetic fields
cannot make electrons
flow.
Electric fields
Electric fields are created by separation of charge. If electrons and their atoms are
separated by a distance, there is an electric field between the two points of the charges. In
fact, there is an electric field between every charged particle. If the charges are same (e.g.
both negative) the field repels the two particles. If the charges are opposite (e.g. positive-
negative) then the field causes attraction between the particles.
Electric fields can be generated in a number of ways, including batteries, other power
supplies, and biological currents. Electric fields cause electrons to move if connected by a
conducting material.
Magnetic fields
ECN 3/e - 16
Magnetic fields can cause electrons to flow, especially if the magnetic field is changing.
The magnetic field induces movement of electrons in a conducting material. Magnetic
fields are generated in a variety of ways, including spin-orientation of atoms in a material
and even electron movement, itself. Movement of electrons in a conducting material
causes a magnetic field which may in turn affect the flow of electrons. This is discussed
further later.The electric and magnetic fields produce an electromotive force (EMF),
which is just what it sound like. A force which causes motion of electrons. The EMF has
a vector which represents the strength and magnitude of the force.
Current flow
Current flow is the movement of charge. Electrons are mobile whereas nuclear protons
are not. So in an electronic circuit, current is the flow of negatively-charged electrons.
Electrons flow from one atom to another, essentially jumping from one partially-filled
orbital to the next.
Electrons are constantly moving, but the movement is random. This is not considered to
be current because there is no net flow of electrons in one direction, i.e. there is no net
vector of electron flow.
If an electric or magnetic field is applied to a conductor, current flows. The field causes
electrons to flow in a unified direction so that there is a defined vector of electron flow.
The unit quantity of charge could be the electron (-1) but other atomic particles have
different valences, including sodium (+1) and calcium (+2). The standard unit of charge
is the coulomb. One coulomb is equivalent to 6.24 x 1018 positive or negative charges.
Current is movement of charge. The standard unit of current measurement is the ampere
or amp. One amp is equal to one coulomb of charge passing trough a point in a conductor
each second.
Or:
charge
Current =
time
or:
Q
I=
t
Circuit theory
A circuit is a closed loop or series of loops composed of circuit elements. The connectors
are conductors, usually wire or ether conductor on a printed circuit board. The circuit
elements can be:
• Resistor
• Capacitor
• Inductor
• Transistor
• Source of electromotive force
Most other circuit elements are components made from these elements.
Circuit elements
Resistors
A resistor opposes the flow of electrons. Functionally, the resistor turns energy associated
with the electrons into heat. Physically, the resistor dissipates energy imparted to the
electrons by the electric field into heat. This is more than a semantic difference. Electrons
leaving the negative pole of a battery are flowing down the potential gradient established
by the chemical reaction in the battery. The electrons will flow quickly down the
conductor, approaching the speed
of light. However, the resistor
dissipates some of the imparted
energy into heat, so that there is
less energy associated with the
electrons. Therefore, fewer
electrons make the journey. The
amount of current is reduced.
Resistance is measured in ohms,
after the nineteenth-century
German physicist Georg Ohm. If a
volt meter is placed across the
terminals of the resistor, the
recordable voltage difference
Figure 2-3: Theory of resistors. reflects the energy dissipated by
A: A conductor connects the terminals of a battery. Current
will flow from the positive terminal to the negative terminal
the resistor. This is the voltage
through the conductor. Electrons are flowing in the opposite drop. This is a distinction to the
direction. The magnitude of curent will quickly destroy the voltage source which would be
battery. the battery or power supply. The
B: A resistor is inserted into the conductor circuit. This voltage drop is zero when current
reduces the current flow by an amount which is proportional
is flowing. If no current is
ECN 3/e - 18
Capacitors
Capacitors are thin plates of conductor separated by a nonconductor. Therefore, there is
no direct conductive continuity between the terminals of the capacitor. Electrons cannot
flow directly through a capacitor because of this separation of conductors, however, there
is an apparent current, called a capacitative current. Electrons go from the negative
terminal of the battery and onto one plate of the capacitor. The accumulation of electrons
on the plate produces an electric field which repels electrons on the nearby plate of the
capacitor. Electrons flow off this plate, through the conducting wire and into the positive
terminal of the power supply. The electrons entering and leaving the capacitor are not the
same electrons, but they seem so. Therefore, this is a capacitative current.
dV
I =C
dt
Inductors
Induction is the effect of magnetic
fields on charged particles. When
electrons pass through a wire, there
is a magnetic field which is oriented
radially around the wire. The
orientation of the magnetic field is
governed by the right-hand rule, a
law which is often remembered from
Figure 2-5: Magnetic fields created by conductors and physics although the application is
inductors.
Current flow through a conductor produces a magnetic
forgotten. With the fingers of the
field surrounding the conductor. When the conducting right hand closed and the thumb
wire is coiled the induced magnetic fields sum to extended, current flow along the axis
of the thumb produces a magnetic field which is oriened as the curled closed fingers of
the hand.
An inductor is a
circuit element
which consists of a
coil of wire. There
may be a metal core
in the center of the
coil. An inductor
stores energy as a
magnetic field. This
is similar to a
capacitor which
stores energy by
separation of charge,
although the effects
on the circuits are
very different. As
Figure 2-7: Semiconductor theory. current flows
A: Atomic structure of semiconductor material. Silicon (left) is a through the wire,
nonconductor. Doping of this material with a conducting element results
in either available electrons for flow (center, N-type), or available empty
some of the energy
electron orbitals which can temporarily host a flowing electron (“hole” or of the electrons is
P-type). used to generate the
B: A piece of semiconductor material as part of a circuit can conduct magnetic field.
electric current, though not as well as a conductor. When the power
C: Two semiconductor pieces are placed together without a battery.
Some of the spare electrone of the N-type semiconductor move to
supply is turned off,
occupy partially-filled orbitals of the P-type semiconductor. This is the magnetic field
analogous to the diffusion potential of biological membranes. collapses, imparting
some EMF to the
electrons. The causes current to flow for a time after the power supply has been turned
off.
Semiconductors
Semiconductors are substances which semi-conduct. They conduct better than non-
conductors but less well than conductors. In general semiconductors are made from non-
conducting material which has a small amount of conducting material mixed in. This
ECN 3/e - 21
mixing process is termed doping and when the non-conductor has the conducting material
in it, it is doped.
Arsenic has a single unpaired electron which is loosely held. This electron can easily be
donated to a nearby partly-filled orbital leaving the arsenic atom with a charge of +1, due
to the loss of electron. Since loosely-held electron is negatively charged, this
semiconductor is negatively doped or N-doped.
Gallium has an empty orbital in an otherwise filled outer shell of electrons. Gallium can
easily accept an electron to fill its outer orbitals, giving it a charge of -1 with the extra
electron. Since the extra space can accept a negatively charged electron, this
semiconductor is positively doped or P-doped.
Current can easily flow through a semiconductor, though less well than through a
conductor. When only one types of doped material is used, current can flow in both
directions. The magic begins when N-doped and P-doped materials are placed together.
Diodes
Diodes are composed of N-doped and P-doped material placed together. One terminal of
the diode is connected to the N-doped material and the other terminal is connected to the
P-doped material.
ECN 3/e - 22
The region of interface between the N-doped and P-doped materials is of special interest.
Loosely held electrons from the N-doped side move across the junction to the P-doped
side thereby occupying the empty spots in the electron orbitals. This situation is not
electrically neutral but is atomically quite stable; the N-doped material has a positive
charge because of the electrons being donated. The P-doped material has a negative
charge because of the acceptance of electrons. The migration of electrons continues until
a charge is built up across this junction which is sufficient to halt the further migration of
electrons. This electrical equilibrium is similar to the diffusion potential of excitable
tissues, where potassium passes through the membrane down its chemical gradient until
sufficient electric charge is built up across the membrane to stop further migration.
A diode takes advantage of the junction effect between semiconductors to allow current
movement in only one direction. If battery is connected to the diode with positive
terminal attached to the P-side and the negative terminal attached to the N-side, then the
diffusion potential at the NP junction is abolished and current can easily pass through the
diode. Since current is pass, this diode is forward-biased.
On the other hand, if the positive terminal is attached to the N-doped side and the
negative terminal is attached to the P-doped side, the junction potential is enhanced. All
of the loosely-held electrons near the junction have passed from the N-side to the P-side.
Virtually all of the empty electron orbitals on the P-side at the junction are filled with
those loosely-held electrons. Therefore, there are no more loosely-held electrons nor
available holes (empty electron orbitals) for electrons to pass; current does not flow.
Since current cannot flow in the desired direction, this diode is reverse-biased.
positive potential. This device is called a rectifier, and is the most important use of
diodes.
Transistors
The term transistor is derived from the words transfer and resistor, since the transistor
controls the transference of energy across a resistance. Transistors have replaced tubes of
older equipment. While we used to describe transistor in comparison with tube function,
the frame of reference of the readers has changed such that this discussion is of little
value. A better analogy might be an automobile.
If one crosses town using a bike, the speed of the bike is dependent on the force applied
by the feet to the pedals. Although there is some mechanical advantage to the bike
including gear ratios and the ability to coast, in general, all of the energy used to move
the bike and body is supplied by the feet. If the same journey is made using an
automobile, the power to move across town is provided by the engine, but the speed of
the engine is controlled by a foot on the accelerator pedal. The speed of the car is
proportional to the power supplied by the engine which in turn is proportional to the
pressure of the foot on the pedal. The power of the foot is essential for the journey, but
the foot muscles had help, i.e. the engine.
This analogy defines the difference between a passive circuit and an active circuit. A
passive circuit has all of the energy to the system supplied by the applied signal voltage
(as for the bike). The active circuit uses an internal energy supply so that the output of the
circuit can be greater than the input signal voltage (as for the car).
What does this have to do with transistors? The transistor is the essential element in the
car analogy. The applied signal voltage is the foot on the accelerator pedal. The engine is
the device power supply. The current through the output impedance is the power
delivered to the wheels. The output voltage is directly proportional to the signal voltage,
though greatly amplified.
The amplification for most transistors is 9x, but for simplicity of math, we will consider it
to be 10x. If three amplifiers are placed in series, the amplification is 1000x. For most
display systems, the amplification may need to be in the millions.
Circuit laws
There are three basic laws governing circuits which should be discussed. The laws are not
used in daily electrophysiologic practice, but are important for understanding the
electrical behavior of circuits. The laws are:
• Ohm’s law
• Kirchoff’s current law
• Kirchoff’s voltage law
ECN 3/e - 24
Ohm's law
Ohm's law describes the
relationship between voltage
(V), resistance (R), and
current (I) for a resistive
circuit. Figure 2-10 shows a
simple resistive circuit with a
power supply connected to a
resistor. For this circuit, the
supply voltage is equal to the
current multiplied by the
resistance.
V = I×R
Figure 2-10: Ohm’s law. Where voltage is measured in
A: Simple resistive circuit with a battery in series with a resistor.
Current flows through the resistance which is proportional to the
volts, current in amps, and
voltage and inversely proportional to the resistance of the resistor. resistance in ohms.
B: Direct relationship between current and applied voltage.
C: Inverse relationship between current and resistance. Ohm’s law: The applied
voltage is equal to the current
multiplied by the resistance of the circuit.
The implications of Ohm's law are evident from permutations of the law:
The formula may not seem intuitively obvious, but the relationship becomes clearer when
these permutations of the formula are considered. First, one can expect that an increase in
voltage would increase the current passing through a fixed resistance. Also, it is easy to
imagine that increasing the resistance would decrease the current flow produced by a
fixed voltage.
The best analogy to this circuit is a pail of water. The pail is partly filled with water and
there is a small hole in the bottom of the pail. The deeper the water, the more pressure
there is on the water to flow out of the hole and the more water runs out. Likewise,
making the hole smaller decreases the amount of water which can exit the pail. In this
analogy, the water passing through the bottom of the pail is the current, the depth of the
water represents the voltage, and size of the hole represents the resistance – with a
smaller hole having higher resistance to flow.
ECN 3/e - 25
For any node, the sum of the currents flowing into a node is equal to the current flowing
out of a node. Since current flowing out of a node can be considered to be of opposite
sign to current which is flowing into a node, then, the statement becomes:
Kirchoff's current law: The sum of currents flowing into and out of a node is zero.
ΣI=0
Kirchoff's voltage law
Kirchoff's voltage law is not as intuitively obvious the current law. This law states that
for any circuit loop, the sum of the voltage sources equals the sum of the voltage drops.
Voltage drops are circuit elements which reduce EMF to the electrons, and for this
example are only resistors. A voltage drop is measured by placing a volt meter across a
ECN 3/e - 26
resistor, but can be calculated from Ohm's law by multiplying the current times the
resistance of the resistor (V = I x R).
Kirchoff’s voltage law: For any circuit loop, the sum of the voltages sources and voltage
drops is equal to zero.
Consider one of the circuit loops in Figure 2-11. If the power supply voltage is increased,
the current increases since the resistance is fixed, from Ohm’s law. Increased current
results in increased voltage drop across the resistor which is equal to the increased
voltage of the power supply.
Circuit properties
Ohm's and Kirchoff's laws can be used to derive other circuit laws, but the mathematical
derivations are unimportant for clinical use. A basic understanding of the principles is
essential.
Resistors
Basic resistor theory was presented
previously, resistor theory
continues with discussion of
resistors in series, resistors in
parallel, and finally, the RC circuit.
Series resistors
Rtotal = ∑ Ri
This would be expected from Kirchoff's voltage law, since the power supplied to the
circuit (VS) equals the sum of the voltage drops across the resistors (VR1 and VR2). Or:
VS = VR1 + VR2
I is the same for the power supply and both resistors since there is only one pathway for
current to flow. Therefore, I drops out leaving:
RTotal = R1 + R2
or:
RTotal = Σ Ri
Parallel resistors
Parallel resistances are conceptually somewhat more complex. First, we will discuss the
concept, then the mathematics.
Figure 2-12 shows the simplest circuit involving two resistors in parallel. The total
resistance of the two resistors is less than each of the individual resistances. An analogy
would be the pail half-full of water previously discussed. If one tiny hole is punched into
the bottom of the pail, the water will run out, albeit slowly through this high resistance. If
a second hole of the same size is punched through the bottom of the pail, each hole has
high resistance, but there are now two avenues for flow of water, so the total resistance is
less. The same applies to our circuit. There are two avenues for electrons to flow through
resistors to the positive terminal, so the total resistance is less than the resistance of each
resistor. How much less?
To simplify the math, let's consider conductance. Conductance (G) is the reciprocal of
resistance. A material of lower resistance conducts better. Therefore:
Back to the pail analogy - each hole has a conductance which describes the flow of water
for a specific depth of the column of water (analogous to voltage). The total conductance
of the pail is the sum of the conductances of the individual holes. Or:
GTotal = G1 + G2
or:
GTotal = ∑ Gi
ECN 3/e - 28
or:
1 1
=∑
RTotal Ri
Capacitors
Capacitors can be placed in series
and parallel, just as resistors, but
because of the electrical properties,
the formulae are reversed! For
simplicity, parallel capacitors will
be discussed first.
Parallel capacitors
C total = C1 + C 2
or:
C total = ∑ C i
Series capacitors
Capacitors in series have a complex interaction. The charge developed from charging of
one capacitor affects the time to charge the second capacitor. In practice, the effect is
ECN 3/e - 29
affects both capacitors reciprocally. The math is more complex than for resistances, but
the result is similar. For 2 capacitors in series, the reciprocal of the total capacitance is
equal to the sum of the reciprocals of the individual capacitances.
or:
1 1 1
= +
CTotal C1 C 2
or:
1 1
=∑
CTotal Ci
Filters
All biological signa s can be broken down into fundamental frequencies, with each
frequency having its own intensity. Display of the intensities of all frequencies is a power
spectrum. In clinical neurophysiology, e usually are interested in signals of a particular
frequency range, termed the band width. the band width differs for different studies. The
band width is determined by filters – devices which alter the frequency composition of
the signal. There are three basic types of filter:
The LFF is sometimes called the high-pass filter, but this old terminology should be
discarded. The LFF filter-out the lower frequencies. The HFF (also formerly called the
low-pass filter) filters out high frequencies. Particular settings of the filters determines
which exact frequencies are affected.
The 60-Hz filter uses tandem LFF and HFF to attenuate frequencies around 60 Hz.
Unfortunately, the 60-Hz is not perfect, and frequencies close to this are attenuated, as
well. Therefore, the appearance of the biological signal may be altered by use of the 60
Hz filter. This is especially true for EEG.
Resistor-Capacitor circuits
The simplest filter is not used in most neurodiagnostic equipment but discussion is
helpful for understanding the concept. A resistor and a capacitor in series forms the
resistor-capacitor or RC circuit. This is shown in Figure 2-14.
ECN 3/e - 30
The resistor (R) and Capacitor (C) are in series with a power source which is a signal
voltage (Vs). The effects of a sudden change in voltage on current is shown in the figure.
Immediately after the voltage is turned on, current (I) begins to flow between the
terminals of the resistor. This current causes a build-up of charge on the capacitor as
capacitative current flows. The charge on the capacitor opposes flow of further current,
ultimately stopping current flow when the voltage across the capacitor is equal and
opposite to the signal voltage.
The time constant (TC) represents the amount of time required for the capacitor of an RC
circuit to charge or discharge to 1/e of the plateau, where e is the natural base,
approximlately 2.7. This value is independent of applied voltage, as long as the voltage is
constant after a change.
Frequency response
All filters result in a particular frequency response. The filters do not cut off all
frequencies outside of the set range. There is a fall off in response to that frequencies far
below the LFF setting are reduced much more than frequencies a little below the LFF
setting. The same holds for frequencies above the setting of the HFF. Figure 2-15 shows
frequency response ffor a number of filter settings.
Most equipment
comes with default
settings for filters
which are reasonable.
However, these
default values should
be checked, since
there are occasionally
programming errors at
the factory. We have
already experienced
this on one of our
machines; the default
settings for median
sensory NCS were far
different from that for
Figure 2-15: Frequency response of filters any other sensory
A: Effect of time constant on current (I), voltage across the resistor (VR), NCS, and were clearly
and voltag across the capacitor (VC) of an RC circuit. wrong!
B: Frequency-response curves with the RC circuit with differint time
constant. Increasing time constant pushes the curves toward the lower
It is instructive to
frequencies. The decline in response to low frequencies is VR. The decline
in response to high frequencies is VC. record EEG and EMG
and play with the filter
settings. This is best done using digital EEG. where the same epoch of activity, slow, fast,
and spike, can be replayed with different recording parameters. Of course, remember to
return the settings to their correct values before the next diagnostic study!
called the serial port, although the performance is nothing like the ADC chips available in
modern equipment.
Details of the function of ADC devices are presented on the CD, since the function is
unimportant for routine interpretation. Suffice it to say that the ADC device samples the
voltage of the signal voltage at predetermined times. This results in an array of digital
values, where each point has three values – channel ID, time, and voltage. These data can
then be manipulated with simple calculations.
Calculations
The digital data can be manipulated for a variety of purposes. Some of these are:
• Averaging
• Spike detection
• Digital filtering
Digital filtering is the most common application. Calculations performed on the digital
data can remove unwanted frequencies. This can change the appearance of the recording,
so the commonly help impression that frequency and phase distortion does not occur with
digital filtering is unfortunately not true.
Averaging is used especially for EP studies, although sensory NCS also employs
averaging when there is difficulty in eliciting a good measureable response.
Spike detection is used predominantly for patients with suspected epilepsy. Calculations
performed on the digital EEG data can tell whether there has been sharp activity
superimposed on an otherwise normal background.
Needle electrodes
Needle electrodes are used almost exclusively for EMG. While needle elecrodes have
been used for EEG in the past, this application has been largely abandoned. For EMG, the
needle is inserted into the muscle. The voltage changes of nearby fibers cause a very
small amount of current to flow through a coaxial electrode and through is leads to the
amplifier before completing the circuit through the reference electrode and tissue.
Although the electrode impedance is fairly small, the amplifier input impedance is large,
so the actual amount of current flowing into the amplifier is small.
Patient-equipment interface
The electrode-amplifier interface is
a key part of the patient-equipment
interface. Figure 2-16 shows the
interface as a circuit representation.
Vs is the signal voltage and is
generated by the neurons. The
acctive elecctrode is the top lead of
Vs while the reference electrode is
the bottom lead. Re is the
resistance of the electrode attached
to the patient. This is not a resistor
but rather is a produce of the
Figure 2-16: Electrode-amplifier interface
A: Diagram of the electrode-amplifier interface. VS is the electrical properties of the
signal voltage from the body. Re is the electrode resistance. electrode attached to the skin. Rin
Rin is the input resistane of the amplifier. The signal is the input resistance of the
voltage is equal to the sum of voltages across the eletrode amplifier. Again, this is not a fixed
(Re) and the input reistance of the amplifier (Rin). resistor, but rather an effect of the
B: Same diagram as in A, but a small capacitance (C) is
inserted into the circuit. This capacitance is created by electrical circuitry.
proximity of the electrode leads and eletrodes.
The voltage passed to the power
amplifiers and display is the
voltage drop seen across the input resistance of the amplifier, Rin. Therefore, anything
which alters voltage across this resistance alters the measurement, distorting the
recording. This can occur in the following circumstances:
ECN 3/e - 34
The derivation for the formula is on the CD, but the following relationship can be shown:
Vin Rin
=
Vs Re + Rin
This means that the ratio of the voltage seen by the amplifier to the voltage generated by
the signal source is dependent on the ratio of the input resistance to total resistance of the
circuit. Therefore, high electrode resistance causes a higher voltage drop across the
electrode thereby reduccing available voltage to the amplifier.
Amplifiers
Amplifiers use transistors which were discussed previously. The concept is that a
controlling signal voltage is increased in size by the amplifier. Ganging of multiple
ECN 3/e - 35
amplifiers is needed in order to produce the voltage needed to drive the display devices or
the analog-to-digital converters of the computers – these devices work with fractions of a
volt rather than millivolts.
Unequal electrode impedances can change the voltage seen by the input impedances of
the amplifiers, as described above. Therefore, the ability of the differential amplifier to
eliminate common mode is degraded by unequal electrode impedances.
ECN 3/e - 36
Part II:
Electroencephalography
Chapter 3: EEG Basics
Chapter 4: Normal EEG
Chapter 5: Abnormal EEG
Chapter 6: Neonatal EEG
Chapter 7: Special Studies in EEG
ECN-CD: Table of Contents
ECN 3/e - 37
Cortical potentials
Most of the cortical efferents have large cell bodies oriented perpendicularly to the
cortical surface. While the cortex has a convoluted surface, the scalp EEG electrodes see
best the electrical activity arising from the regions which are relatively parallel to the
scalp. Inhibitory and excitatory inputs to these large efferent neurons produce substantial
current which sums to become the scalp EEG. The dendritic arborization is in the
superficial layers of the cortex while the soma and axon hillock are in the deeper layers.
This creates a vertical columnar organization of the cortex.
Scalp potentials
Alpha rhythm
The alpha rhythm is usually seen in normal, relaxed individuals who are awake with their
eyes closed. It is approximately 10 Hz in adults with the maximum voltage originating
from the occipital electrodes, O1 and O2. The term alpha rhythm is used by some
physiologists to signify any posterior dominant rhythm regardless of frequency, but this
is an improper use off the term.
In children, the dominant posterior rhythm is slower and frequency and may not attain the
minimal 8.5 Hz until 12 years of age. Slower frequencies in a 12-year-old would be
interpreted as abnormal and would most likely indicate a diffuse encephalopathy or, if
unilateral, suggest a structural lesion.
ECN 3/e - 40
The posterior dominant rhythm is suppressed by eye opening and promptly returns when
the eyes are closed. This reactivity of the posterior alpha rhythm should be routinely
tested during EEG recording. The posterior rhythm is suppressed if the patient is tense
during the recording. The lack of posterior rhythm should not be interpreted as abnormal
in this situation. Other EEG features that suggest a tense state include frequent eye blinks
and muscle artifact in frontal and temporal leads.
The amplitude of the posterior rhythm is 15-50 µV in young adults. Older individuals
often have lower amplitude, but the frequency is the same. Low amplitude should not be
interpreted as abnormal if the frequency composition is normal. Slowing of the posterior
dominant rhythm is not a normal part of aging. Amplitude asymmetries of the dominant
rhythm are common. The amplitude is usually higher from the non-dominant hemisphere,
but the difference should not exceed 50%.
A prominent alpha rhythm can be recorded during anesthesia and coma, but the
distribution is different from the normal posterior alpha rhythm. In anesthesia and coma,
the alpha rhythm is generalized with an anterior predominance. This alpha activity is
invariant and monotonous, lacking the usual modulation in frequency and amplitude of
an occipital alpha. The appearance of alpha coma in a patient signifies a poor prognosis
for good neurologic recovery.
Beta rhythms
EEG activity with frequencies faster than 13 Hz occurs in all individuals but is usually of
low amplitude and often overlooked in favor of slower frequencies during wakefulness
and sleep. Beta activity is normally distributed maximally over the frontal and central
regions. A low-amplitude high-frequency beta is especially prominent during normal
sleep in infants and children and is enhanced by several sedatives, especially barbiturates
and benzodiazepines. In some children, the beta activity is so prominent as to dominate
the record.
People with hyperthyroidism may accelerate their posterior rhythm from 10-14 Hz or
more. This is technically in the beta range, but the rhythm continues to react like an
occipital, awake, resting rhythm, and should be considered no different than the alpha
rhythm in this context.
Theta rhythms
EEG activity with a frequency between 4 Hz and 8 Hz is seen in normal drowsiness and
sleep in adults. Young children may show theta activity during the waking state, which
ECN 3/e - 41
Posterior slow waves of youth may be in the theta or delta range. Theta activity in the
temporal region in older individuals has been ascribed to vascular disease. While the
significance of temporal theta is controversial, we suspect that that it is not part of normal
aging. We suggest commenting on the presence of temporal theta in the body of the
report and interpreting it as a mild abnormality.
Delta rhythms
Delta activity is not normally recorded in the awake adult but is a prominent feature of
sleep and becomes increasingly abundant during the progress from stage 2 to stage 4
sleep. Focal polymorphic delta activity may be recorded over localized regions of
cerebral damage. Intermittent rhythmic delta activity is recorded when there is
dysfunction of the relays between the deep gray matter and cortex. The activity has a
frontal predominance in adults and is called frontal intermittent rhythmic delta activity,
while in children the activity has an occipital predominance and is called occipital
intermittent rhythmic delta activity or posterior intermittent rhythmic delta activity.
Generation of
Epileptiform activity
Epileptiform activity is
generated when
depolarization of the cortex
results in synchronous
activation of many neurons. It
is conceptually attractive to
equate action potentials with
EEG spikes, but action
potentials occur normally.
The abnormalities in
epileptiform activity are the
repetitive nature of the
Figure 3-2: Topographic distribution of a spike focus. discharge and the degree of
Sample topographic distribution of a spike focus. The negative end synchrony. By virtue of the
is in the right frontal region whereas the positive pole is in the nature of scalp EEG
occipital region. The dipole is often oriented perpendicularly to the recordings, synchronous
cortical surface so that the positive end of the dipole is not seen on
the cortex. activation of many neurons is
required for generation of
both normal and abnormal rhythms.
Epileptiform activity consists mainly of spikes and sharp waves. While epileptiform slow
activity and suppression does occur, this is much less common. Spikes have a duration of
less than 70 msec while sharp waves have a duration of 70-200 msec.
Usually, the negative end of the epileptiform dipole points to the cortical surface.
Therefore, most recorded spikes are scalp-negative. The distribution of the negativity
across the scalp surface is the field. The field can be represented as a topographic map,
as shown in figure 3-2.
Paroxysmal depolarization shifts (PDS) are extracellular field potentials which are waves
of depolarization followed by repolarization. High-amplitude afferent input to the cortex
produces depolarization of cortical neurons sufficient to trigger repetitive action
potentials, which in turn contribute to the potential recorded at the surface. Repolarization
due to inactivation of interneurons is followed by a brief period of hyperpolarization.
Technical aspects of
EEG
EEG equipment
EEG has changed greatly in
recent years. For many years,
machines were fairly standard,
Figure 3-4: Diagram of analog vs digital EEG recording. analog machines with paper
A: Flow chart of the function of analog EEG. The initial 10x display. Sixteen channels was the
amplifier is to elevate the magnitude of the signal so that norm, though additional channels
there is sufficient signal voltage to reduce apparent artfat and
signal degradation. helped with localization. Most of
B: Flow chart of the function of digital EEG. The analog-to- the recent advance has been in
digital converter, montage generator, and digital filters are the conversion to digital
part of the computer which is the core of the machine. The machines. These usually use a
display may also be integral to the computer, but can also be CRT display, although there are a
an external, secondary high-resolution display.
few older machines which would
print on paper. Now, paper print
is mainly used for only selected portions of the record rather than the entire record.
EEG equipment is either analog or digital. Analog machines are commonly in use,
although digital machines are of increased popularity. Analog machines show a direct
representation of the cerebral signal amplified many times to be displayed on paper.
Analog filters change the frequency components of the displayed signal prior to writing
on paper. Digital EEG machines produce display output which looks similar to the output
of an analog machine, but the signal handling is very different. Cerebral signal is
amplified then digitized, as discussed in Chapter 2. Digital filters alter the frequency
components of the signal. The display is usually on a computer monitor although the
results can be written on paper. In the interest of conservation of paper and ease of
interpretation, electronic display is preferable to paper display. Electronic display allows
for alteration of filters, gain, and montage “on-the-fly”.
Digital interpretation should be used only with caution and not as a substitute for visual
interpretation. Quantitative differences may be significant mathematically yet
meaningless clinically. Similarly, mathematical analysis may miss findings of great
clinical importance. No hardware and software combination is close to threatening human
interpretive abilities.
Display montage: Most digital machines record the potential from each electrode so that
the montage can be altered by the reader. A spike seen on a longitudinal bipolar montage
can then be viewed on a transverse bipolar montage. This can help localization, but the
reader has to realize that multiple views of the same event do not represent multiple
occurrences of the event.
Gain: High-amplitude transients can peg the pens of an analog display, so some
information is lost. Digital machines allow the gain to be changed, so that high and low-
amplitude epochs will not be lost.
5). Inertial distortion is due to the fact that the pens have mass and pressure on the paper.
Therefore, it takes more energy to get the pen moving vertically than it does to keep it
moving. Although the paper is moving horizontally, getting it to move vertically requires
more energy. The other extreme of inertial distortion is potential overshoot of the pen
when it is moving vertically. In the absence of some corrective maneuvers, the pen
overshoots its vertical target, termed overshoot. Mechanisms to compensate for these
mechanical distortions on paper recordings include electrical compensation of the
amplifiers and adjusting pen pressure. Damping is the term used to describe the
compensatory mechanisms for inertial distortion. Digital machines are immune to these
mechanical distortions.
Number of channels
The effective minimum number of channels for routine recording is 8, however, most
neurophysiologists require 16 and 21 adds additional channels to montages which can
greatly help localization. ICU monitoring for patients in status epilepticus can be
performed with 4 channels, however, a baseline recording with a complete set of
electrodes should be performed for evaluation of the discharge and background.
The number of factors considered in making a decision is hugs, however, there are a few
factors which deserve special mention.
Cost: Cost of the machine is a small part of the expense of an EEG lab. The space and
technical help are costs which dwarf the cost of the EEG machine. Therefore, it makes
little sense to skimp on machine features for the sake of small amount of money.
Portability: At least one of the hospital EEG machines must be portable so that ICU, ER,
and OR recordings can be made. There are special requirements for equipment which will
be brought into the OR, but most modern machines meet these requirements. This has to
specified, however, if use in the OR is at all anticipated.
Reading stations: This is a moot point for paper EEG, but digital EEG can only be
interpreted at workstations. Number, location, and availability of workstations should be
ECN 3/e - 46
considered. For example, in one of our hospitals, EEG data is retained on a drive in the
machine, so when a portable study is being performed in the OR or ICU while
disconnected from the network, EEGs performed on that machine cannot be read. This is
not acceptable.
Electrodes
Surface electrodes
Surface electrodes are used for almost all routine EEG applications. They are disks which
are fixed to the skin in a variety of ways. Electrode gel forms a malleable connections
between the rigid disk and the skin. The electrode is secured on the scalp usually by
pressing a gauze pad onto the gel-electrode combo. This can hold for a short time
providing the patient is cooperative so that the head does not move. This electrode system
does not work well if the patient moves or if the patient has stiff hair or other impediment
to fixation of the electrode. Collodion fixation is a much more secure method of securing
the electrodes on the scalp, although it does require special equipment, adequate
ventilation, and more technician time. We usually use collodion preferentially in our
laboratory even for routine outpatient studies.
1. Locate the positions for electrodes using the 10-20 Electrode Placement System
(explained later)
2. Separate strands of hair over the electrode positions using the wooden end of a
cotton-tipped applicator.
3. Clean dead skin and dirt from the region with an agent such as Omni-Prep using
the cotton-tipped applicator.
4. Scoop some gel into the electrode.
5. Place the electrode in position over the skin.
6. Put a 2"x2" gauze pad over the electrode and push it firmly onto the head,
providing a seal which prevents the electrode from falling off the scalp.
1. Prepare the head at the electrode positions as mentioned for electrode gel.
2. Place the electrode on the scalp
3. Place a piece of gauze soaked with collodion over the electrode
4. Use compressed air top dry the collodion
5. Insert a blunt-tipped needle into the up and scrape the skin to lower electrode
impedance
6. Inject electrolyte into the cup of the electrode using the blunt-tipped needle.
Removal of electrodes is easy for the gel fixation. The gauze pads are pulled off then the
electrodes gently pulled off, tilting them to release the vacuum effect which holds them
on. Then, the gel left on the scalp can be largely removed by rubbing with a warm, wet
ECN 3/e - 47
wash cloth. After the patient washes the hair that evening, all traces of the recording are
gone.
Collodion is more difficult to remove. First, the collodion is softened by use of acetone,
then the areas cleaned as above. The degree of washing required in greater both
immediately by the technician and later by he patient. Some patients object to the acetone
smell more than any other part of the procedure.
Each method has its advantages. Collodion provides a more secure attachment and is
more suitable for long-term recordings. Electrode gel is easier to apply and remove and is
suitable for most routine office and hospital recordings.
Needle electrodes
Needle electrodes offer no advantages over conventional surface electrodes and should
not be used for routine studies unless recording cannot be accomplished any other way.
The risk of infection to the patient and technician is unacceptably high.
Sphenoidal electrodes
Sphenoidal electrodes are used to record activity from the temporal lobe which would not
show on scalp recordings. The electrodes are inserted percutaneously adjacent to the
zygoma until they reach the base of the skull. Sphenoidal electrodes should only be used
by physicians trained in their insertion and experienced in interpretation of the recorded
potentials.
Subdural strip electrodes are used to evaluate patients for epilepsy surgery. The strips are
placed during surgery through burr holes. The strips allow for a detailed map of the
recorded electrical activity. Subdural strip electrodes should only be used by physicians
trained and experienced in placement and interpretation, and only as part of a
comprehensive epilepsy intervention program.
Depth electrodes
Depth electrodes are used to localize seizure foci for surgery. A depth electrode consists
of an array of electrodes on a single barrel which is inserted into the brain, usually in the
temporal lobe. Only trained and experienced epileptologists should use depth electrodes.
Electrode position
Electrodes are named according to their position within the 10-20 system, with the first
part of the name referring to the region and the second part referring to the area within
the region. For regions, the following assignments have been made:
• F = frontal
• C = Central
• P = Parietal
• T = Temporal
• O = Occipital
• A = Auricular (ear)
• Fp = Frontopolar
The second part of the electrode name indicates the exact location. Numbers refer to pre-
arranged location, and the reader has to know the 10-20 system to know the difference
between T3 and T5, for example. Odd numbers are on the left side of the brain, even
numbers are on the right. Within a region, lower numbers are more anterior and medial to
higher numbers. Midline electrodes are designated with "z" rather than a number.
1. Measure the distance from the nasion to inion across the vertex. Mark a line 50%
of this distance a the top of the head.
2. Measure the distance between the preauricular points, just in front of the ear.
Mark a line at 50% of this distance, at the top of the head. The intersection of this
line with that of step 1 is Cz.
3. Lay the measuring tape from nasion to inion through Cz. Mark 10% of this
distance above the nasion for Fpz and above the inion for Oz. Fz is 20% of this
distance above Fpz. Pz is 20% of this distance above Oz.
4. Lay the tape between the preauricular points through Cz. T3 is 10% of this
distance above the left preauricular point and T4 is 10% of this distance above the
right preauricular point. C3 is 20% of this distance above T3, and C4 is 20% of
this distance above T4.
5. Lay the tape from Fpz to Oz through T3. FP1 is 10% of this distance from Fpz, F7
is 20% of this distance posterior to Fp1. O1 is 10% of this distance anterior to Oz,
and T5 is 20% of this distance anterior to O1. Measure in the same manner for
Fp2, F8, O2, and T6 over the right hemisphere.
6. Lay the tape from Fp1 to O1 through C3. F3 is half the distance between Fp1 and
C3. P3 is half the distance between C3 and O1. Repeat for the right side, with the
tape from Fp2 to O2 through C4. F4 is half the distance between Fp2 and C4, and
P4 is half the distance between C4 and O2.
7. Lay the tape from F7 to F8 through Fz., F3 and F4 to ensure that the distance
between the electrodes is equal. Then lay the tape from T5 to T6 through Pz, P3,
and P4 to ensure equal interelectrode distances.
Abbreviations for special electrodes and less standardized, and may differ between
laboratories. Sp usually indicates sphenoidal electrode and Naso usually indicates
nasopharyngeal electrode. Odd numbers are left-sided and even numbers are right-sided.
Subdural strip and depth electrodes are also named using letters and numbers where the
letter indicates the array and the number indicates which electrode in he array.
Montages
The sequence of electrodes being recorded at one time is called a montage. All montages
fall into one of two categories: bipolar or referential. Referential means that the reference
for each electrode is in common with the other electrodes. The reference could be a single
non-cephalic electrode, ipsilateral ear, or perhaps even Cz, though this is not a useful
reference because of it’s electrical activity. Bipolar montage means that the reference
electrode for one channel is the active electrode for the next channel. Bipolar montages
are particularly useful for visual analysis of focal cerebral activity such as spikes and
sharp waves. For example, in the longitudinal bipolar (LB) montage, the first four
channels are shown in table 3-2.
For all channels,, negativity at the active electrode produces an upward deflection of the
pen on the paper. Negativity at the reference produces a downward deflection. The
Guidelines recommends the following principles in designing montages:
The recommended montages for routine use in adults are shown in Table 3-3, and
diagrammed in Figure 3-7. Additional channels, when available, are used for monitoring
other biological functions, such as ECG, eye movements, respirations, and EMG.
ECN 3/e - 51
Routine EEG
EEG recordings are usually stored on paper or disk. There is a face sheet which is
attached to the paper record, if one exists, or is separate, in the case of digital recordings.
For all recordings, the face sheet has all of the basic identifying information:
• Name
• Age
• Identification number of the patient
• Index number of the recording
• Reason for the study
• Name of the technologist
• Current medications
• Time of the last seizure, if appropriate
• Technical summary, including activation methods and artifacts
• Technicians observations, including regions of particular concern or interest.
• Time and date of the recording
• Ordering clinician
• Sedative medication used
Digital recordings should be identified on the face sheet, with the index number of the
disk and the format of storage.
The face sheet should be filled out completely before the patient leaves the lab. If the
technician sees a finding of immediate clinical concern, the neurophysiologist should be
called immediately.
Calibration
Two phases of calibration are used prior to each study - square-wave calibration and
biological calibration.
Square-wave calibration
A square-wave pulse is delivered from a wave-form generator into each amplifier input.
This pulse is 50 µV in amplitude and alternated on and off at 1 second intervals. The
wave does not appear a precise square wave because of the effects of the preset default
filters. Figure 3-8 shows a sample recording.
The low-frequency filter (LFF) transforms the plateau of the square wave into an
exponential decay. The high-frequency filter (HFF) slightly rounds off the peak of the
calibration. For educational purposes, try several HFF and LFF settings during the
calibration test to see the effects of filter changes on the record. It is also instructive to
change filter settings during recording of EEG activity at a time which would not
interfere with clinical interpretation.
The time constant (TC) of the LFF can be measured from the square-wave calibration
page. TC is equal to the time it takes for a potential to fall to 37% of peak value. To eye-
ball the value, this is approximately one third of peak value.
It is difficult to estimate the HFF setting from the square wave calibration, however,
neurophysiologists should have an idea of what the peak should look like. If the HFF is
ECN 3/e - 53
set too low, there will be a slow roll-off on the peak of the calibration pulse. If the HFF is
set too high, the wave will appear too peaked and may even show overshoot, as if there is
too little pen damping.
Biological calibration
Biological calibration, of Biocal, assesses the response of the amplifiers, filters, and the
recording apparatus on a complex biological signal. Electrodes Fp1 and O2 are connected
to all of the amplifier inputs. The recordings from all of the channels should be identical.
Pen pressure and damping are issues which apply only to paper recordings, digital
recordings are immune to these causes of distortion. Mechanical writing instruments have
two inherent limitations: inertia and friction. Even when the filters are set properly, the
frequency response may be inaccurate because of these mechanical factors. The physical
mass of the pen produces inertia that slows its response time to sudden changes of signal
voltage. Inertia is partially compensated for by control mechanisms in the pen drive
mechanism. Friction is also compensated by EEG machine electronics, but excessive
pressure of the pen on paper results in a sluggish response. This is the main reason that
ECN 3/e - 54
The same inertia that inhibits pen movement also promotes excessive pen movement,
overshoot. Overshoot is minimized by the pen control mechanism, termed damping.
When damping is not sufficient, normal waveforms may look like spike discharges.
These effects are minimized with proper setting of pen pressure and damping. The
manuals provided with EEG machines give instructions on setting of damping and pen
pressure.
Sensitivity
The recording sensitivity is initially set at 7 µV/mm and subsequently adjusted depending
on the amplitude of the EEG activity. Movement artifact and other non-cerebral
transients may exceed maximal pen excursion, but electrocerebral activity may not.
Important waveforms may be missed when the sensitivity is set too low.
For children, the sensitivity is often reduced to 10-15 µV/mm because EEG amplitude is
high in both the awake and sleep states. The elderly often have low-voltage EEG activity,
and increased sensitivity is required.
Studies performed for the determination of brain death are started at 7 µV/mm but the
sensitivity is always increased to 2 µV/mm.
Duration
Filters
• LFF = 1 Hz
• HFF = 70 Hz
The LFF of 1 Hz corresponds to a TC of 0.16 sec. If the LFF is set higher than 1 Hz,
there will be attenuation and distortion of some slow waves. Slow waves have an
increased number of phases and are composed of faster frequencies. Technicians should
be discouraged from turning up the LFF especially when there is an abundance of slow
activity. If the HFF is set too low, fast activity is blunted, and spikes and sharp waves
may be impossible to identify.
ECN 3/e - 55
The 60-Hz filter should not be needed in most laboratories. Three ways to minimize 60-
Hz activity are:
Shielding of the room is desirable but usually not essential and cannot completely abolish
artifact induced by strong electromagnetic fields. Studies in a special care unit usually
require use of the 60-Hz filter. Sources of artifact include ventilators, intravenous
infusion pumps, air beds, heating or cooling blankets, and monitoring equipment.
Activation methods
The performance and interpretation of records obtained with activation methods are
discussed in Chapter 4. Hyperventilation, photic stimulation, and sleep may activate
epileptiform activity. After an initial period of recording in the relaxed, wakeful state, the
patient is asked to hyperventilate for 3 minutes. If absence seizures are suspected, the
patient is asked to hyperventilate for 5 minutes. Hyperventilation is not performed in
elderly individuals or in patients with advanced atherosclerotic disease because of
concern for vasoconstriction with resultant cardiac or cerebral hypoperfusion.
Photic stimulation is performed on older children and adults of all ages. Photic
stimulation of sleeping infants is probably of limited clinical value.
Electrode impedance
Electrode impedance should be at least 100 ohms and no more than 5 kohm. Lower
impedances cannot be obtained with scalp electrodes without there being some improper
connection between the electrodes, such as smeared gel between adjacent electrodes.
Higher electrode impedances can create impedance mismatch which can then degrade the
rejection of artifact by bipolar recordings.
Electrical interference is common in EEG laboratories, but since most recordings are
bipolar, even referential recordings, the noise is typically similar in conformation and
amplitude in the leads. Since the bipolar recording setup subtracts the voltage of the
reference from the active signal voltage, the noise cancels out. Signal which is in
ECN 3/e - 56
common to the two inputs is called common mode. This rejection can be described by the
common mode rejection ratio. This represents the ability of an amplifier to reject signal
in common to two inputs of the the amplifier.
Electrode caps are commonly used in facilities which seldom perform EEGs. The
technical staff usually does not perform EEGs as their primary duty, so the caps offer
efficiency and relatively consistent electrode placement.
Technicians at outside facilities are seldom as well trained and supervised as technicians
in the home EEG laboratory. Therefore, there is greater opportunity for technical error.
ECN 3/e - 57
Many outside facilities use electrode caps which fit snugly on the head. There are several
problems with electrode caps:
EEG laboratory
EEG can be performed in almost any part of the hospital, but routine studies are
performed in the EEG lab, where electrical and acoustic noise is minimal, and all
ancillary equipment is available. Many labs have the patient in a separate room from the
recording equipment and technician, but this is not usually necessary for routine daytime
recordings.
The EEG laboratory should conform to standard guidelines for electrical safety, including
3-prong plugs with proper grounding. Shielding is usually not needed with proper
grounding, good technique, and bipolar montages. However, if the EEG laboratory is
forced to be in an electrically noisy location, electrical shielding may be necessary. If the
laboratory is in an acoustically noisy location, acoustic insulation may be necessary.
These factors greatly increase the cost of the laboratory.
Report
The EEG report should be clear and concise, and ideally be no longer than one page.
Sample recordings are seldom attached to EEG reports, as they may be for EP and EMG
reports. The three fundamental components to the report are:
• Report header
• Description of the record
• Interpretation
Report header
Patient information:
• Name of the patient
• Date of birth
• Sex
• Identification number
Laboratory information:
• Name of laboratory
• Address and phone number of laboratory
Study information:
ECN 3/e - 58
• Date of study
• Reason for study
• Referring physician
• Interpreting physician
The description of the record should be complete without containing exhaustive detail.
Among the information included in the description of the record is:
• State and state changes - wake, drowsy, sleep - and description o the findings at each
level.
• Background rhythms - posterior dominant rhythm, frontal rhythms, temporal theta
• Focal abnormalities - asymmetries, focal slowing, focal sharp activity
• Epileptiform activity - spikes and sharp waves whether focal or generalized, and
regional prominence
• Artifacts which interfere with interpretation of the record, and impression whether the
study is interpretable
Interpretation
The interpretation is the most important part of the report. While there is considerable
room for personal preferences, the following information must be provided:
• Normal or abnormal
• How the recording is abnormal
• Clinical implications of the findings
Or:
The clinical interpretation should take into account the information provided by the
ordering physician. For example, if the clinical history is of partial complex seizures and
the EEG shows a spike focus, the interpretation might be:
Abnormal study because of a spike focus in the right anterior temporal region. This is
consistent with a partial seizure disorder.
On the other hand, if the clinical history is of a behavioral disorder, the impression might
be:
ECN 3/e - 59
Abnormal study consistent with a sharp-wave focus in the left posterior-temporal region.
This could be consistent with a seizure disorder , but patients without seizures can
manifest this pattern.
Record keeping
Printed EEG reports should be kept for the duration of the practice, although there are no
clear published recommendations on this issue. Patients with seizures may live entire
lives with multiple studies performed over time, and a comparison can be extremely
helpful.
The paper EEG record cannot be kept forever, but at least parts of the record should be
kept for at least 2 years. Microfilming can greatly improve record keeping of paper
records. Digital recording has allowed for virtually indefinite storage of the complete
record, and will be the predominate method of storage in the future. Optical disk, CD,
and DVD recordings have a duration and reliability which certainly exceed routine
storage requirements.
State and local regulations may have additional requirements for record keeping.
ECN 3/e - 60
Anterior cerebral
activity
The background
activity of the frontal
region is composed
predominantly of low-
voltage fast activity
with superimposed eye
movement artifact.
Rhythmic beta may be
recorded from the
frontal and central
regions, especially
when sedatives are
used. Drug enhanced
beta is more commonly
Figure 4-1: Normal waking EEG. seen after
Normal EEG recorded using the left parasaggital portion of the
longitudinal bipolar montage. There is a posterior dominant rhythm of 9- benzodiazepine and
10 Hz which attenuates to eye opening (signified as “OE!”) barbiturate sedation
than following chloral
hydrate.
Theta and delta activity are not prominent in the normal awake adult EEG. However,
digital EEG signal analysis shows a small amount of bihemispheric theta in most patients.
In the awake state with the eyes closed, the predominant rhythm from the occipital region
ECN 3/e - 61
is about 10 Hz. The range of normal is 8.5 to 11 Hz, with frequencies slower than 8.5 Hz
being abnormal, usually indicating dementia or encephalopathy. There is some oscillation
of this rhythm, but it still has a very regular appearance.
The posterior dominant rhythm is usually symmetric, but asymmetries of up to 25% are
frequently seen. Asymmetry should not be interpreted as abnormal unless it is at least
50%. The amplitude from the left hemisphere is often lower than the right, so this should
be considered when interpreting amplitude asymmetries.
Increasing age results in the posterior dominant rhythm being lower amplitude and less
well organized. While slowing of the background is common in the elderly, it is still
abnormal. We use a rigid cutoff of 8 Hz; any frequency less than this is abnormal at any
adult age.
Tense state can suppress the posterior dominant rhythm even with the eyes closed, so
relaxation is important to recording an adequate response.
Vertex waves may be seen at stage 1B, but are more a sign of stage 2 sleep.
Differentiation of stages 1A and 1B is not important for routine EEG.
Sleep
Sleep patterns
POSTs
Positive occipital sharp transients of sleep (POSTS) are surface positive potentials with
maxima at O1 and O2. They may occur as single waves or in trains. They resemble
lambda waves except that they are present only in the sleeping state, whereas lambda
waves are seen only in the waking state and with the eyes open.
POSTS may be related to replay of visual information during sleep, but this hypothesis is
not universally accepted. They are not seen in patients who are blind or are severely
visually impaired.
POSTS are not a consistent feature of sleep and have no diagnostic significance, unless
they are seen only from one side. This asymmetry is unlikely to be the only sign of focal
abnormality in an EEG record.
ECN 3/e - 62
BETS
Benign epileptiform
transients of sleep (BETS)
are very small spike-like
potentials that occur in the
temporal regions during
drowsiness and light sleep.
They are less than 50 µV
with a duration of less than
15 ms. Also called small
sharp spikes, BETS are
differentiated from
epileptiform pikes by their
small amplitude, short
duration, lack of a slow
wave, and normal EEG
background.
Figure 4-2: POSTs
Positive occipital shart transients of sleep (POSTs) recorded using the
left parasaggital ortion of the longitudinal bipolar montage. The Vertex waves
upward deflections in the last channel are due to positive potentials
from the O1 electrode. Vertex waves are surface-
negative potentials with a
maximum amplitude on
either side of the midline (C3 and C4). They are most common in stage 2 sleep and often
appear at times of partial arousal. The vertex waves are a biphasic sharp wave with an
initial negative deflection followed by a positive deflection. The sharp wave may be
followed by a slow wave or a spindle, the latter being termed a K complex.
The vertex waves can be asymmetric, especially in children, and high amplitude in
younger patients. The asymmetry should be abnormal if more than 25% and if consistent
between the hemispheres. Vertex waves first appear at approximately 8 weeks of age. In
children, vertex waves that appear in trains may be mistaken for seizure activity.
Sleep spindles
Sleep spindles are rhythmic 11-14 Hz waves whose duration is typically 1-2 sec, with a
minimum of 0.5sec, and whose amplitude is at least 25 µV. They are most prominent in
the central regions during stage 2 sleep. Unlike vertex waves, the maximum amplitude of
sleep spindles is typically seen lateral to the midline (C3 and C4). Asymmetry in the
abundance of sleep spindles is normal unless sleep spindles fail to appear from one
hemispheres.
Sleep stages
Stage 1
Stage 1 is drowsiness, and is divided into stage 1A and 1B. Stage 1A is characterized by
attenuation of the posterior dominant rhythm and some change in field of distribution so
that it is seen more anteriorally than in the fully awake state.
Stage 1B is characterized by progressive loss of the posterior dominant rhythm, with less
than 20% of the background composed of the alpha rhythm. Theta activity becomes more
prominent. Vertex waves may be present in stage 1B, however, this is more commonly
seen in stage 2. Differentiation between stages 1A and 1B is not important for routine
EEG.
Stage 2
Stage 2 sleep is characterized by sleep spindles, vertex waves, increased theta, and the
appearance of delta. However, less than 20% of the record contains delta. Since some
vertex waves can be seen in stage 1B, the primary differentiating feature is the
appearance of sleep spindles.
Stage 3
Stage 4
REM
Typically, REM sleep follows after progression from sleep stages 1 through 4.
Progression from drowsiness to REM sleep without passing through other stages, termed
REM-onset sleep, occurs in the following conditions:
• Narcolepsy
• After sleep deprivation
• After alcohol or drug-induced REM deprivation sleep
The features of REM sleep which distinguish it from drowsiness are the following:
• Rapid and chaotic eye movements (drowsiness has slow roving eye movements)
• Hypotonia on submental EMG
• Irregular respiratory rate
Sequence of sleep
stages
Patients progress
from relaxed
wakefulness through
stage 1A and 1B into
stage 2 sleep. During
a routine office
EEG, deeper stages
of sleep are seldom
seen. In fact, when
seen they may be
Figure 4-3: Maturation of the posterior dominant rhythm. misinterpreted as
The posterior dominant rhythm gradually increases during childhood years. abnormal rhythms.
Reaching adult frequencies by 4-5 years of age. However, there are other Progression into
differences which differentiate a child’s EEG from an adult EEG. slow-wave sleep,
ECN 3/e - 65
EEG in children
Waking EEG
Maturation of the posterior rhythm
The posterior dominant rhythm in the awake infant is approximately 4 Hz. The rhythm
becomes faster with age, reaching the normal adult frequency of approximately 10 Hz by
10 years. This maturation is shown in figure 4-3. The amplitude gradually increases, such
that by age 10 years the alpha is often in the range of 50-100 µV. In adults, the alpha
amplitude gradually declines with increasing age.
Slow waves of youth are seen predominantly in the waking state and occasionally in light
drowsiness, stage 1. They are in the delta range and are superimposed on the normal
posterior dominant rhythm. Slow waves of youth may be differentiated from pathologic
slow waves by the otherwise normal background and their reactivity to eye opening. slow
waves of youth decrease with increasing age and are not seen after the age of 30 years.
Sleep EEG
Sleep promotes some epileptiform activity and is used as an activation method. However,
the morphology of the epileptiform discharges may look different in sleep than they look
in the waking state. Sleep patterns differ in children and adults.
ECN 3/e - 66
The posterior dominant rhythm of relaxed waking state attenuates with drowsiness. As
the alpha disappears, theta activity appears from both hemispheres. This is commonly
referred to as stage 1 sleep.
Vertex waves
Vertex waves in children appear at about the age of 5 months. They are initially blunted
in morphology. By 2 years of age the vertex waves are prominent, sharp in configuration,
and high in amplitude. Clusters of vertex waves may be mistaken for epileptiform
activity, especially since they can occur in trains and can be asymmetric, and can have
appear apart from the midline.
Sleep spindles
Sleep spindles first make their appearance at about the age of 2 months. They are often
prolonged and asymmetric. By 2 years, the sleep spindles have an adult appearance.
Normal
transients and
variants
<< Table 4-3: Normal
transients and variants. >>
Lambda waves
Lambda waves are normal
Figures 4-5: Sleep spindles.
positive waves that appear
Sleep spindles during stage 2 sleep. This is a portion of an average
over the occipital region
reference montage. The spindles in this example are independent on
the two sides of the head. when the patient is looking
at a picture or pattern.
Lambda waves indicate
visual exploration and are blocked by eye closure. They are called lambda waves because
of their resemblance to the Greek lowercase letter lambda (λ).
Mu rhythm
Mu rhythm is not a common feature of normal EEGs. It is a run of negative wicket-
shaped spikes with an approximate frequency of 10 Hz and a duration from less than 1
second to many seconds. The negativity is maximal in the rolandic regions, mainly C3
and C4. Mu has the appearance of a centrally located alpha rhythm but is usually slightly
ECN 3/e - 67
faster than the patient’s alpha rhythm. It is blocked by movement of the contralateral
extremity; this is the key to identification. In fact, Mu can be blocked by merely thinking
about limb movement. The technician should ask the patient to move a limb to verify the
waveform.
Wicket spikes
Wicket spikes are sharply contoured waves that are most prominent from the temporal
regions during drowsiness and light sleep. They are differentiated from true spikes by:
The 14-and-6 positive spike rhythm has been associated with a number of pathological
conditions, but the association is generally weak and the incidence is likely not different
from that in the general population. Therefore, this should be considered a normal
variant, if the background is otherwise normal. Metabolic encephalopathies, such as
hepatic encephalopathy, may have an increase incidence of this pattern, but the
background is abnormal, as well.
Encephalopathy rarely manifests as a slow background rhythm without any other sign of
abnormality; the organization and frequency composition of the more anterior regions is
key to differentiation.
Rhythmic mid-temporal
theta of drowsiness is
differentiated from
seizure activity by:
• Normal background
before and after the
rhythm
• Absence of a
progressive frequency
change which typifies
seizure activity
Figure 4-6: Rhythmic temporal theta of drowsiness. • Presence in
Thythmic temporal theta of drowsiness sees with maximum in the drowsiness but not in
anterior temporal region. This is the left lateral portion of the sleep.
longitudinal bipolar montage.
This pattern is
considered a normal variant, but has been described in association with structural lesions.
SREDA
Subclinical rhythmic electrographic discharges of adults (SREDA) is rhythmic sharp
waves which evolve into a rhythmic theta pattern. It has an abrupt onset and changes in
conformation and frequency throughout the discharge. SREDA is seen in older patients
during the awake state. SREDA can easily be confused with seizure activity. Some
important differentiating features include:
ECN 3/e - 69
Mittens
Mittens are seen only in sleep and consist of a partially fused sleep spindle and vertex
wave. The last wave of the spindle is superimposed on the rising phase of the vertex
wave. The voltage summation gives the last spindle wave a faster appearance, which
simulates a spike. The name is derived from the appearance of a mitten, the thumb being
the fused spindle wave and the hand portion being the slow component of the vertex
wave.
Mittens are normal, but can be confused with spike-wave patterns. Careful examination
can usually dissect out the components comprising the mitten. While older literature
correlated mittens to psychiatric disorders and tardive dyskinesia, this should still be
considered a normal pattern.
Non-cerebral potentials
Eye movement
Eye movement artifact is seen in the
anterior leads in virtually all records.
The eye is polarized with the cornea
positive relative to the fundus.
Therefore, when the eye rotates to look
down, the leads over the frontal region
are close to the negative end of the
ocular dipole. This effect is most
prominent for Fp1, Fp2, F3, and F4. The
reverse is true with upward gaze, as the
frontal leads are close to the positive
end of the dipole. With lateral gaze, the
Figure 4-7: Eye lead placement electrodes most affected are F7 and F8.
A: Eye lead placement and sample reccordings with For example, with left gaze, F7 becomes
vertical eye movements. This positioning helps with
differentiation of eye movements from frontal slow more positive while F8 becomes more
activity. negative.
B: Alternative method of recording eye movements.
This allows for differentiation of horizontal fromDifferentiating eye movement artifact
vertical eye movements. from electrocerebral activity is usually
not difficult. First, eye movements have
a stereotypic pattern that looks different from most abnormal frontal slow activity, which
is usually more polymorphic. The onset of the slow waves caused by eye movement is
rapid with a slower decay. Also, eye movement waveforms are typically superimposed on
a normal low-voltage, high-frequency background. Abnormal frontal delta activity is
usually associated with increased theta activity and reduced beta activity in the frontal
ECN 3/e - 70
Eye leads can be placed in several ways. The two most common methods are shown in
Figure 4-7. We use the method shown in Figure 4-7A. Electrodes are placed above and
lateral to the right eye and below and lateral to the left eye. These electrodes are
referenced to an average or ear electrode. With upward gaze, the positive end of the
dipole rotates toward the right lead but away from the left. This cause pen deflections of
opposite polarity in the recording. With left gaze, the positive cornea rotates toward the
left lead but away from the right. Again, the pen deflections will be in opposite
directions. These electrode derivations will detect slow activity in the frontal lobes, but
this slow activity will not reverse between the two sides. Therefore, in these channels,
slow activity that is opposite in polarity is of ocular origin, while slow activity that is of
the same polarity on both sides is most likely of cerebral origin.
The method shown in Figure 4-7 of eye movement detection allows for precise
determination of direction of gaze. Vertical gaze can be distinguished from horizontal
gaze. This is seldom of interest on routine EEG testing.
Eye opening
Eye opening results in not only the eye movement artifact described above, with the eye
moving up and down with the blink, but also alters the posterior rhythm. There is prompt
attenuation of the posterior dominant rhythm with replacement with low-voltage fast
activity.
Muscle artifact
EMG activity is
frequent contaminant
of EEG recordings. It
is most prominent in
Figure 4-8: Muscle artifact
Marked muscle artifact in a tense patient. This is the left parasaggital the awake state and is
portion of the longitudinal bipolar montage. characterized by fast,
short-duration spikes
in the temporal and frontal regions. Amplitude is approximately 50 µV. EMG activity is
ECN 3/e - 71
• EMG activity is very fast. In fact, the predominant frequency is much faster than the
HFF setting of 70 Hz.
• EMG activity is not followed by a slow wave.
• EMG is most prominent in the waking, tense state, and disappears with relaxed
wakefulness and sleep. In contrast, epileptiform activity is often best seen in
drowsiness and sleep.
• EMG activity can often be attenuated by asking the patient to open his mouth. The
masseter and pterygoid muscles do not contribute much to surface EMG activity, so
this relaxation promotes decreased activation of scalp muscles and a reduction in
scalp muscles without adding further artifact from jaw muscles.
Glossokinetic artifact
The tongue is negative at the tip in comparison to the base, forming a dipole similar to the
eye. Movement of the tongue is usually seen in the waking state as delta activity which
can be mistaken for pathological delta activity. Talking can exacerbate the glossokinetic
artifact. The artifact disappears with drowsiness and light sleep.
Movement artifact
Movement produces artifact in two ways: movement of the electrode leads and
perturbation of the electrode-gel interface.
Electrode leads
The flow of electrons through wires creates a weak magnetic field. Magnetic fields can in
turn influence the flow of electrons through conductors – they induce current. This
interaction between current and magnetic fields is termed inductance and is the physical
basis for inductors, discussed previously. Inductance is an important mediator of noise.
Electrode leads are unshielded and therefore susceptible to the effects of ambient
magnetic fields. These magnetic fields are created by current in nearby power lines and
equipment. Magnetic fields then induce current to flow through the electrode leads. This
mechanism is termed stray inductance because the inductance is not intentional within
the machinery. The induced current flows through the electrode leads just as signal
ECN 3/e - 72
current does. The amplifier cannot distinguish between signal and noise, so both are
amplified. Since line power is 60 Hz, this created 60 Hz interference.
The differential amplifier will reject much of the 60-Hz interference; however, the
rejection is incomplete under the following conditions:
A diffusion potential is established at the interface between the electrode and the gel,
which is caused by movement of ions between the two substances. The concept is similar
to the creation of a resting membrane potential in neuronal membranes. When there is
movement of the head, this junction is disturbed and the junction potential discharges,
injecting current through the electrode into the input amplifier. This is interpreted by the
amplifier as a voltage pulse.
ECN 3/e - 73
The appearance of this artifact is a brief spike followed by a gradual decay to baseline.
During the pulse, the responsiveness of the amplifier may be reduced, since the amount
of current low from the artifact can be sufficient to overload the input amplifier. If there
is no further movement, a new equilibrium is established.
Machine artifact
Machine artifact is usually a high-frequency artifact which may be at 60-Hz, the
frequency of line power. This is true if the origin is a power supply, motor, or other AC-
powered component. However, since most electronic circuits are DC-powered, machine
artifact is more commonly due to DC-powered motors or resonant electric activity.
Motors are contained in IV pumps, ventilators, hospital beds, among other devices.
Figure 4-9 shows the machine artifact created by an air bed. Briefly unplugging the bed
results in a hard surface for the patient, but a tremendous improvement in the EEG
recording.
Electric artifact can be generated by equipment and perceived as machine noise. Among
the devices with inherent resonance are radios.
Electrocardiogram artifact
ECG is recordable from
some patients, especially
young children and with
the use of referential
montages. Increased
inter-electrode distance
predisposes to increased
ECG artifact. If there is a
question as to whether
there is ECG artifact, a
separate channel with
cardiac leads can be
placed and the relative
timing of the brain spikes
and ECG compared.
Figure 4-10: ECG artifact
ECG artifact is seen in this left parasaggital portion of the ear reference Pulse artifact
montage. Ear reference montages is much more likely to show ECG
artifact than bipolar or average refferencce montages.
Pulse artifact is related to
ECG and is due to
movement of an electrode over an artery. Pulsation of the vessel moves the electrode and
lead slightly, altering the diffusion potential of the electrode-gel connection and altering
the electromagnetic interaction between adjacent electrode leads. This produces an
irregular delta wave which can be localized to the lead overlying the artery. Identification
ECN 3/e - 74
• Localizing the artifact usually to one electrode; cerebral structural lesions would
produce slowing recorded by more than one electrode.
• Timing of the pulse to the ECG; there is a brief delay between the ECG QRS complex
and the pulse but the timing should be consistent.
• Otherwise normal background; patients with polymorphic delta activity due to a
structural lesion will typically have an abnormal background, with theta activity and
absence of a well-developed posterior dominant rhythm.
If there is still doubt, then examination of the scalp can show the pulsing vessel.
Movement of the electrode off of the vessel can confirm the artifact, though this is
seldom necessary.
Activation
methods
Activation methods are
used to bring out
epileptiform discharges in
patients with suspected
seizure disorders.
Hyperventilation and
photic stimulation are
routinely used. Sleep is
considered by some to be
an activation method, but
is considered separately,
above, since it is a normal
Figure 4-11: Photic evoked potential state change rather than an
Left medial portion of the longitudinal bipolar montage, with channel
1 showing the stimuli. Flash at 5/sec produces an evoked potential in applied method.
the fourth channel, due to activity in the occcipital lead. The upgoing
potential in this bipolar montage indicates positivity at the O1 Photic stimulation
electrode. The positivity is delayed from the stimulus by about 100
msec, indicating that this is an evoked potential rather than a photic
response. Photic stimulation is
repetitive flashes of light
delivered at different rates.
The flash stimulus can evoke epileptiform discharges at certain flash rates. Photic
stimulation is more likely to evoke epileptiform discharges in patients with generalized
epilepsies than in patients with focal epilepsies.
Methods
The stimulating protocols are pre-programmed into most EEG machines. General
guidelines for performing photic stimulation include the following:
ECN 3/e - 75
If the discharge is
activated as a specified
frequency, the technician
should repeat that
frequency at the
completion of the photic
stimulation routine.
Normal responses to
photic stimulation include
the visual evoked
response, the driving
response, and the
photomyoclonic response.
Driving response: A driving response is seen at flash frequencies of 7/sec and greater.
The two responses look alike but are distinguished by their temporal relation to the
stimulus. The visual evoked response occurs approximately 100 ms after the stimulus,
and the driving response is exactly time-locked to the stimulus.
The absence of a visual evoked response or a driving response is not abnormal unless it is
well developed on one side and absent on the other. Such asymmetry suggests an
abnormality affecting either the projections from the lateral geniculate to the cortex or the
calcarine cortex, itself.
Photomyoclonic response
ECN 3/e - 76
Photoconvulsive response
Hyperventilation
Hyperventilation is usually used
to activate the three-per-second
spike and wave discharge of
primary generalized epilepsy. In
some patients, discharges are
seen only during
hyperventilation. The patient is
asked to mouth-breathe deeply
for approximately 3 minutes. If
Figure 4-13: Photoconvulsive response
Photoconvulsive response, usually produces by a speccific there is suspicion of absence
range of photicc frequencies. The discharge is not time- seizures, the patient should
locked to the stimulus and typically last longer than the hyperventilate for 5 minutes.
stimulus. Left medial portion of the LB montage.
The normal response to
hyperventilation is generalized slowing of the background activity in the theta range in
both hemispheres. Absence of slowing is nor abnormal and depends on effort, age, and
time from last meal, Children show more slowing than adults with hyperventilation.
Hypoglycemia may augment slowing.
ECN 3/e - 77
Slowing
Slow activity is frequently normal, such as theta during drowsiness and delta during
sleep. However, focal delta during the waking state or theta for a posterior dominant
rhythm in the waking state is clearly abnormal. Slowing can be divided into three
classifications:
• Generalized slowing
• Regional slowing
• Focal slowing
Generalized slowing includes the slow activity which would characterize encephalopathy,
with slowing of the posterior dominant rhythm, disorganization of the rhythm, and
excessive theta activity anteriorally.
Focal slowing is usually indicative of a structural lesion, and includes focal theta activity
and polymorphic delta activity.
should not be considered abnormal unless it stands out from the background and is
reproducible. The alpha rhythm may occasionally appear sharp, even though the pointed
component is positive. Positive polarity helps to distinguish this waveform from a spike
potential.
Most spikes and sharp waves in adults are normal unless they are artifacts. Several
normal spike-like potentials can be seen, however. These include:
• Vertex waves
• Occipital lambda waves
• 14- and 6-Hz positive spikes
• Wicket spikes
• BETS – benign epileptiform transients of sleep
• POSTS – positive occipital sharp transients of sleep
• 6-per-second (phantom) spike and wave
These are discussed in detail in Chapter 4, and can sometimes be difficult to distinguish
from epileptiform activity. The relevant discussions should help with this distinction.
The electrophysiologist is not blind to the clinical question when interpreting the EEG.
The best approach is to describe the record, a phase of interpretation which would be
independent of the clinical question. The Impression will then consider the EEG findings
in light of the clinical question. Therefore, detailed clinical information provided to the
electrophysiologist certainly aids clinical utility of the EEG. Most of the time that true
spikes and sharp waves are seen, the clinical correlation is seizures, and in this case, the
impression might read:
“Abnormal study because of epileptiform activity arising from the right temporal region.
This would be consistent with a partial seizure disorder with a focus in the right temporal
lobe.”
However, if the clinical question is not seizures, the clinical interpretation of these
findings is more controversial. We believe that mention should be made of the
epileptiform appearance of the activity but also the disclaimer made that not all patients
with epileptiform activity have seizures. For example, the same EEG pattern discussed
above in a child with behavioral disorder might result in the following impression:
ECN 3/e - 82
Table 5-2: EEG patterns with selected disorders which demonstrate spikes.
Disorder EEG pattern
Rolandic epilepsy Stereotypes spikes, often with triphasic appearance, followed by a slow
wave. Independent discharges with maxima near C3 and C4 Activated
by sleep. Background is otherwise normal.
Occipital epilepsy Unilateral or bilateral high-amplitude occipital spikes which are
increased by light sleep and suppressed by eye opening.
Rapid discharge from the same regions during a seizure.
Absence epilepsy Classic 3-per-second spike-wave or polyspike-wave complex. Increased
by hyperventilation. Less well organized in sleep.
Subacute sclerosing Periodic high-amplitude bursts of sharp waves, often with an irregular
panencephalitis delta wave superimposed. Synchronous between hemispheres. Duration
of 0.5-2.0 sec. Many seconds to a minute or more between bursts.
Creutzfeldt-Jakob Periodic complexes with a frequency of 0.5-2.0/sec. Maximum
anteriorally. Abnormal background, with low-voltage slowing. Periodic
complexes abate during sleep.
Anoxic From normal to isoelectric depending on severity. Common patterns
encephalopathy include burst suppression and a periodic pattern that resembles CJD.
Alpha coma is a diffuse nonreactive alpha activity with a frontocentral
predominance; these latter patterns indicate a poor prognosis.
Herpes simplex Generalized slowing with a frontotemporal predominance. Periodic
encephalitis complex of sharp waves or sharply-contoured slow waves, which appear
irregularly. High-amplitude discharges with a frequency of 0.2-1.0/sec.
Arbovirus Slowing in the theta and delta range with little intra- or interhemispheric
encephalitis synchrony. Few faster frequencies.
Lennox-Gastaut Slow spike-wave complex. Slow background in many patients.
syndrome Increasing disorganization during sleep.
Juvenile myoclonic Generalized polyspike discharges followed by a slow wave. Higher
epilepsy amplitude in the frontal region. Otherwise normal background.
Complex partial Various patterns depending on site of origin of epileptiform activity.
seizures May be unilateral temporal or frontal spikes. Focal slowing or no
discharge recordable with surface electrodes.
Simple partial Midline spikes with a prominent negative phase or biphasic in most
seizures patients. Occasional patients may have a positive prominence on surface
EEG. These patterns correlate with simple motor seizures.
Generalized tonic- Generalized polyspike discharge that often, but not invariably, has a
clonic seizures slow wave. Ictal activity is often high-frequency spikes without obvious
slow waves.
“Abnormal study because of epileptiform activity arising from the right temporal lobe. In
the appropriate clinical situation, this would be supportive of a partial seizure disorder,
however, the presence of this pattern of epileptiform activity does not mean that the
patient is certain to have seizures”.
Abnormal patterns can be focal or generalized, and the clinical implications depend on
the exact pattern and location. However, there are some generalizations:
• Focal slowing – usually suggests a focal structural lesion underlying the scalp
electrodes.
• Focal spikes or sharp waves - can correlate with a focal structural lesion but more
commonly suggests a partial seizure disorder. The type of seizure correlates with
location. Of course, not all sharp activity indicates seizures.
• Diffuse slowing – usually associated with encephalopathy, which can have many
potential causes, including toxic, metabolic, degenerative, and multifocal vascular
disease.
• Diffuse spikes or sharp waves – correlate with a generalized seizure disorder. Cannot
rule-out secondary generalization.
Slowing of the background rhythms is discussed in Slow activity. Excessive fast activity
is usually seen in patients sedated with benzodiazepines – beta activity is prominent
frontally. Theta activity is present in almost all recordings, and can be seen if the gain is
high enough. Theta is not a prominent component of the background in waking adults,
and when it stands out from the baseline is abnormal; there are several potential clinical
correlations, discussed below.
Slow activity
Diffuse slowing
Diffuse slowing can have several presentations. The most common is slowing of the
posterior dominant rhythm in the waking state. Occasionally, slow activity can be
superimposed on an otherwise normal waking background. Identification of abnormal
slow activity in sleeping records is especially challenging.
The normal adult waking EEG consists of mainly fast rhythms. With eyes closed,
rhythms in the alpha range are seen from the posterior regions and faster frequencies are
ECN 3/e - 84
seen from the frontal regions. Slowing of the PDR to less than 8.5 Hz is always abnormal
in adults. Slowing of the posterior dominant rhythm (PDR) is usually seen as the
posterior rhythm in the theta range, e.g. 6-7 Hz. The slow posterior dominant rhythm
differs from the normal faster rhythm in a few ways:
• Slow PDR is less stereotyped than normal PDR, with bumps on the waves
• Slow PDR is less reactive to eye opening than normal PDR, it does not show the
degree of attenuation of normal PDR
• Slow PDR is often associated with theta prominent more forward of the occipital
regions than the normal PDR extending forward of the occipital regions.
The slow PDR is interpreted as being abnormal, but is not specific. Possible causes
include:
• Toxic-metabolic encephalopathy
• Degenerative dementia
• Multifocal vascular disease
The impression when this is the only finding might be: “Abnormal study because of
slowing of the posterior dominant rhythm. This is suggestive of a diffuse encephalopathy,
although it is a nonspecific finding.” Comment might be made about metabolic, toxic,
and dementing causes, depending on the specified clinical question.
Occasionally, the PDR has the appearance of a subharmonic – where there may appear to
be a 5-6 Hz PDR with otherwise normal frequency composition and appearance of the
EEG. This is a normal variant, and should be interpreted as normal. The subharmonic
PDR can be differentiated from slowing of the PDR in the following ways:
• Slowing of the PDR in the 5-6 Hz range should be associated with slowing seen
anteriorally to the occipital lobes, whereas subharmonic PDR has otherwise normal
frequency compositions.
• Slowing of the PDR in the 5-6 Hz range will usually not attenuate completely to eye
opening, whereas subharmonic PDR completely attenuates.
• Slowing of the PDR in the 5-6 Hz range with have an irregular, polymorphic
appearance, whereas subharmonic PDR is regular, and usually notched, so that the
underlying 10 Hz rhythm can be seen.
Theta and delta activity in waking records is usually abnormal. Diffuse slowing is usually
polymorphic delta or irregular theta which is seen from both hemispheres. The slowing
does not typically have the regional concentration of the normal PDR or frontal fast
ECN 3/e - 85
activity. Most of the time, the PDR is slow when there is diffuse slowing, but not always.
Diffuse theta with a temporal prominence can be associated with a PDR still in the alpha
range.
Focal slowing is irregular and composed of delta activity with theta activity
superimposed. Even faster activity is then superimposed on this background. This
irregular appearance is the reason for the term polymorphic delta activity (PDA).
PDA may not be continuous, although this is surprising when considering the pathologic
processes which cause PDA and the physiology behind the EEG patterns. Why should a
fixed structural lesion create a synchronous potential shift which would be episodic? The
PDA often appears on a disorganized EEG background, but the background may actually
be normal.
ECN 3/e - 86
PDA is the most common finding in focal structural lesions such as tumors, contusion,
hemorrhage, infarction, and abscess. The presence of focal spikes or sharp waves without
another disturbance on the background is seldom a sign of a focal parenchymal lesion.
Focal slowing is nonspecific, there are no characteristics that distinguish one cause from
another. Complicated migraine and postictal state may cause focal slowing.
The rhythmic slowing of FIRDA and PIRDA may last for several seconds then disappear
for longer intervals, hence the intermittent nature of the rhythm. The slow activity is
augmented by eye closure or hyperventilation, but attenuated by stimulation or by non-
REM sleep. FIRDA reappears in REM sleep.
A wide variety of lesions can produce IRDA, so the interpretation should indicate the
abnormal nature of the rhythm without implications for localization. There are no major
diagnostic differences between FIRDA and PIRDA. PIRDA is seen occasionally in
children with absence epilepsy. Both FIRDA and PIRDA can be caused by:
ECN 3/e - 87
• midline tumors
• metabolic encephalopathy
• degenerative disorders
• some encephalitides
•
• FIRDA is differentiated from PDA by the latter’s lack of reactivity to the stimulus,
usual unilateral appearance, lack of rhythmicity, and the continuous appearance.
Focal spikes are interpreted if the spike is consistent, has an identifiable field, and cannot
be explained by artifact. A single spike during the course of a recording should not be
ECN 3/e - 88
Focal spikes are associate with partial seizures and the benign epilepsies of childhood.
Partial seizures are divided into simple and complex, based on symptomatology rather
than EEG findings. The benign epilepsies of childhood can manifest as focal and
generalized seizures.
The EEG during a simple partial seizure usually shows prominent spiking over the
involved cortex, although in some patients there may be localized slowing which may
become generalized. A typical pattern might be left central spikes in a patient who
presents with focal seizures affecting the right arm. Occasionally, the sharp component of
the discharge may be subtle or missing. The epileptiform activity may occur in deep
layers of cortex and subcortical structures so that the spike potentials are not projected to
the surface electrodes. Alternatively, there may not be sufficient synchrony to produce a
spike detectable on the surface.
clinically, but this is not always the case. The generalization may occur so quickly that
the focal onset can only be determined by EEG, and not by clinical appearance.
Benign focal epilepsies of childhood are termed benign because they are age-related and
seldom persist into adult life. There are two types: rolandic and occipital.
Rolandic epilepsy
The discharges of rolandic epilepsy are so characteristic in location and pattern that they
are seldom confused with other pathologic activity. Independent central spikes are seen
on an otherwise normal background. This must be differentiated from multifocal spikes,
however.
Occipital epilepsy
ECN 3/e - 90
During the seizure, the EEG shows 2-3/sec spike-wave discharges with predominance in
the occipital region. The interictal discharge may be blocked by photic stimulation or eye
opening.
Focal spikes or sharp waves are occasionally seen in patients with no clinical seizures.
The EEG may have been ordered for some non-epileptic indication. Some of these are
children who are genetic carriers of benign epilepsies, whereas in others there is no
explanation. The interpretation of these records is controversial. Some neurophysiologists
believe that all abnormal sharp activity is potentially epileptogenic and should be
interpreted as such. Unfortunately, this may result in unneeded use of antiepileptic drugs.
Patients should be treated with antiepileptic drugs based on clinical presentation rather
than on EEG findings. The old adage still is valid “Treat the patient, not the EEG.” Of
course, the countering argument is that the patient may have seizures which are not
always clinically identifiable.
Children with behavioral disturbances have been reported to have an increased incidence
of focal sharp waves and spikes. The implication of these findings is controversial. Some
investigators believe that the spikes may have contributed to the behavioral disturbance
by interfering with normal social and intellectual development. Others believe that the
spikes are incidental and should not be treated. The spikes are probably a reflection of
brain dysfunction, which correlates with the behavioral disorder rather than being the
cause of the dysfunction.
Some patients with congenital blindness may exhibit occipital spikes. These should not
be interpreted as epileptiform.
3-per-second spike-wave
Appearance
The 3-per-second spike-wave complex is synchronous from the two hemispheres, with
highest amplitude over the midline frontal region. The lowest amplitudes are in the
ECN 3/e - 92
temporal and occipital regions. The frequency changes slightly during the course of the
discharge, beginning close to 4/sec and declining to 2.5/sec. Immediately following the
discharge, the record quickly returns to normal. The spike component may have a double
spike or polyspike appearance.
The 3-per-second
spike-wave complex
is promoted by
hyperventilation. If
absence epilepsy is
considered, the patient
should be asked to
hyperventilate for 5
minutes instead of the
usual 3 minutes.
Children with absence
seizures become
symptomatic if the
discharge lasts longer
than 5 seconds.
Figure 5-6: Three-per-second spike-wave complex
During the discharge,
The 3-per-second spike-wave complex which is typical of absence the technician should
seizures. The spike is actually a polyspike, and the frequency is not ask the patient a
constant throughout the discharge, with faster frequency during the early question. The patient
portion of the discharge and slowing somewhat later in the discharge. Left with absence seizures
and right medial portions of the LB montage.
often answers after the
discharge. The
question and the
response should be noted on the record.
The 3-per-second discharge is less well organized during sleep than during the waking
state. Its appearance is more polyspike in configuration and the spike-wave interval is
less regular.
The spike component is polyspike in some patients. Patients with this polyspike pattern
are more likely to exhibit myoclonus.
Clinical correlations
The 3-per-second spike-wave pattern correlates well with primary generalized epilepsy, if
the remainder of the recording is normal. Factors which would make the clinical doubt
the diagnosis of primary generalized epilepsy include:
ECN 3/e - 93
Treatment of absence epilepsy often abolishes the interictal discharge. This is different
from most focal epilepsies in which interictal spiking persists despite good seizure
control.
During sleep, the slow spike-wave complex may be continuous. This activity may not
indicate status epilepticus but rather represents activation of the interictal activity with
sleep.
The slow spike wave complex is frequently associated with the Lennox-Gastaut
syndrome. It also has been called petit mal variant, but this term is misleading and should
not be used. In the Lennox-Gastaut syndrome, the slow spike-wave complex is usually an
interictal pattern, but may also be ictal. Since these patients have a mixed seizure
disorder, ictal events may show patterns other than the slow spike-wave complex,. Atonic
seizures are characterized by generalized spikes during the myoclonus followed by the
slow spike-wave pattern during the atonic phase. Atonic seizures are most characteristic
of the Lennox-Gastaut syndrome. Akinetic seizures are characterized by the slow spike-
wave discharge throughout the seizure. Tonic seizures occur in Lennox-Gastaut
syndrome and are characterized by a rapid spike activity or desynchronization rather than
the slow spike-wave complex.
The fast spike-wave complex has a frequency of 4-5/sec and has the appearance of slow
waves with superimposed sharp activity, rather than distinct spike-wave complexes.
Maximal amplitude is in the fronto-central region.
ECN 3/e - 94
Patients have
generalized tonic-clonic
seizures with or without
myoclonus. Absence
seizures are rare. This is
the most common
pattern seen in patients
with idiopathic
generalized tonic-clonic
seizures. The discharge
is not as stereotyped as
the 3-per-second spike-
Figure 5-7: Fast spike-wave complex wave complex, and the
Fast (about 6-per-second) spike-wave complex, which is seen in synchrony is less
primary generalized epilepsies. Left lateral portion of the LB montage.
prominent.
The clinical implications of the frontal and occipital rhythms differ. Frontal
predominance is frequently associated with generalized tonic-clonic seizures, whereas
occipital predominance is not associated with clinical seizures. Hughes (1980) provided
the acronyms WHAM and FOLD. WHAM = waking record, high amplitude, anterior,
males. FOLD = females, occipital, low amplitude, drowsy. WHAM is associated with
seizures and FOLD is not.
The 6-per-second spike-wave pattern is differentiated from the 14-and-6 positive spike
pattern not only by polarity but also by the more widespread distribution and occurrence
in wakefulness. Both rhythms may appear in the same patient. The 6-per-second pattern
is interpreted as abnormal and the different clinical implications should be emphasized in
the report.
Hypsarrhythmia
Periodic patterns
Periodic discharges
usually indicate cortical
damage, and can be due
to stroke, anoxia,
infection, degenerative
disorders, and other
conditions. The periodic
patterns can be focal,
regional, or generalized,
with regional
distribution being the
most common.
Figure 5-8: Hypsarrhythmia
Hypsarrhythmia, seen typically in children with infantile spasms. The Periodic lateralized
high-amplitude bursts with interburst interval is characteristic. Left and
right lateral portion of the LB montage.
epileptiform
discharges
PLEDs are high-amplitude sharp waves that recur at a rate of 0.5-3.0/sec. They are
prominent over one hemisphere or one region. When bilateral, they are independent,
thereby keeping the term lateralized.
PLEDs are a sign of parenchymal destruction and most commonly seen in strokes. Other
important causes include head injury, abscess, encephalitis, hypoxic encephalopathy,
brain tumors, and other focal cerebral lesions. It is impossible to distinguish definitively
between causes on the
basis of waveform. Of
the encephalitides, herpes
simplex most commonly
produces PLEDs. Other
viral infections produce
slowing without PLEDs.
Patients with PLEDs may have myoclonic jerks that are either synchronous with the jerks
or independent. When the jerks are independent, the generator for the myoclonus is
probably deep. Even when they are synchronous, the generator is probably subcortical.
The cortical discharge reflects projections from the deep generator.
HSV encephalitis usually shows PLEDs on EEG during some phase of the illness,
although at other times, there is slowing in the theta and subsequently delta range. The
PLEDs are sharply contoured slow waves with a frequency of 2-4 Hz. The duration of
each wave is often more than 50 msec. This relatively slow frequency of repetition helps
to differentiate PLEDs in herpes encephalitis from the higher frequency discharges of
SSPE.
Neonates with herpes encephalitis may have necrosis that is not confined or even most
prominent in the temporal region. These patients often do not have PLEDs. The EEG
may show a poorly organized background with slow activity in the delta range
predominating.
Anoxic encephalopathy
The background is disorganized with diffuse slowing and suppression. Periodic sharp
waves are often seen and may predominate in the record. They look similar to PLEDs,
except that they are synchronous between the hemispheres. Patients may have myoclonus
associated with the discharges. These probably represent the extreme of the burst
suppression pattern, seen often in patients with anoxic encephalopathy.
Burst-suppression
pattern
The burst-suppression
pattern occurs in
patients with severe
encephalopathies. The
finding is not specific
as to etiology but is
most often seen in
patients with hypoxic-
ischemic damage and
Figure 5-10: Burst-suppression pattern
Burst-suppression pattern, which some people call the suppression-burst in barbiturate coma.
pattern since the bursts occpy a lesser temporal proportion of the recording
than the suppressions. Right medial portion of the LB montage. The burst-suppression
ECN 3/e - 97
pattern in different clinical conditions can look very similar. In fact, the burst suppression
pattern can look similar to the markedly discontinuous pattern of 29-week conceptional
age newborns. Bursts of slow waves with superimposed sharp activity are superimposed
on a very suppressed background. The background is not flat, but rather is very low
voltage, composed of a mixture of frequencies.
EEG in SSPE resembles the burst-suppression pattern. The background is usually more
suppressed with burst suppression than SSPE. The two patterns are more easily
differentiated by clinical presentation. Patients with burst suppression usually have a
known history of hypoxia or severe metabolic derangement. Patients with SSPE have a
typical history of a progressive neurologic disorder with intellectual deterioration and
seizures. SSPE is very rare.
Creutzfeldt-Jakob disease
characterized by low-voltage slowing in the theta and delta range. The periodic
complexes abate in sleep.
Early in the course, the periodic complexes cannot be seen and the only finding may be
focal or generalized slowing. About 10-15% of patients may not show periodic patterns
during their course.
ECN 3/e - 99
• Respirations
• Eye movements
• ECG
Respirations can be rapid and produce an artifact that mimics slow activity on the EEG.
ECG monitoring helps with identification of cardiac and pulse artifact. Eye movement
recordings and submental EMG help to identify wake and sleep states.
Newborns spend most of their time sleeping so sedation is usually not necessary. The
electrodes are placed before the baby is fed. Then the baby usually quickly falls asleep
after feeding. The baby should be aroused later in the study to observe arousal and the
waking state.
Photic stimulation is of little benefit in newborns and is not routinely performed. Driving
responses are not consistent and photoconvulsive discharges are rare at this age.
Hyperventilation cannot be voluntarily performed and is not artificially performed.
Montages
Recommended montages for neonatal EEG are shown in Table 6-1. The most commonly
used montage is the Newborn montage. This is a version of the longitudinal bipolar (LB)
montage with fewer scalp leads; there are two channels between the fronto-polar and
occipital regions rather than four for the adult LB montage. However, since the newborn
head is smaller, the inter-electrode distances are comparable.
ECN 3/e - 100
The montage is commonly changed during an adult recording, but we prefer to select a
single montage for the entirety of a neonatal recording. This allows for easy detection of
state changes. We routinely use the newborn montage.
• Gestational age
• Postnatal age
• Physiologic state
• Reactivity
• Clinical question
Conceptional age is used for interpreting neonatal EEG, Conceptional age is the sum of
gestational age plus postnatal age. Maturation of the EEG has defined stages, which are
presented below.
Physiologic state refers to the wake-sleep cycle. The cycles and terminology are different
in adults and children. This is also discussed below.
in AS is characterized by theta activity with some delta and beta activity superimposed.
During the course of a long sleep, the first AS period is higher amplitude than subsequent
AS periods. The later AS periods have more theta and less delta.
The EEG shows long periods of low-voltage activity punctuated by short bursts of
higher-voltage activity. The bursts are composed of mixed frequencies. Sharply
contoured theta and faster frequencies can give the normal bursts an epileptiform
appearance, but this is a normal pattern. The interburst intervals may last up to 2 minutes,
although intervals of less than 1 minute are more typical. When the bursts first develop,
there is poor synchrony between the hemispheres. With full development, there is good
interhemispheric synchrony.
occipital regions and are best seen in AS. Delta brushes resemble sleep spindles but are
physiologically different. Sleep spindles are not prominent in REM sleep and are minimal
in the occipital regions. Also, the disappearance of delta brush later in development and
the subsequent development of frontal spindles argue against a common physiologic
substrate.
The EEG during QS is still discontinuous, although the intervals of quiescence are shorter
and less pronounced. Delta brushes are still present, and the spindle component is of
higher frequency. Slow waves in the delta range are seen in posterior leads. AS is still
discontinuous. Chin EMG is reduced during AS but is not a reliable indicator of state.
Multifocal sharp transients appear at this stage, occurring in the wake and sleep states.
They are differentiated from pathological spikes by their widespread distribution and lack
of repetitive discharge.
The EEG in QS (non-REM sleep) is still discontinuous but the interburst intervals are
progressively shorter with increasing age. The burst-interburst time ratios are 1:2 and 1:3.
AS (REM sleep) is virtually continuous, with delta predominating posteriorally and theta
and faster frequencies anteriorally. For the first time, EMG becomes a reliable indicator
of state, with low-amplitude in REM sleep.
Multifocal sharp
transients are less
prominent and are
replaced by frontal
sharp transients.
These are of higher
voltage than
multifocal sharp
transients. The EEG
is more reactive to
external stimuli than
at earlier ages.
Figure 6-2: Trace-alternant pattern.
Newborns at term often have a discontinuous pattern termed trace-anternant.
Stimulation causes
The discontinuity is less prominent than in premature infants. attenuation of the
background and
frequently is
followed by a change in state.
Term infants have good differentiation between REM sleep, non-REM sleep, and
wakefulness. During non-REM sleep, the discontinuous pattern now has a burst-
interburst ratio of about 1:1. This is the mature trace alternant (TA) pattern. Non-REM
sleep may be characterized by a continuous slow wave pattern rather than TA. This
pattern is occasionally misinterpreted as encephalopathy in neonatal EEG.
Frontal sharp transients are less prominent but may be seen until 2 months of age. Delta
brushes are absent.
Abnormal patterns
Abnormal neonatal EEG patterns fall into at least one of the following types:
• Abnormality of maturation
• Epileptiform activity
• Background abnormality
Abnormalitites of maturation
Dysmature means that the EEG pattern is not appropriate for the conceptional age. For
example, a discontinuous pattern with an interburst interval of 1 minute is normal in a
preemie o 29 weeks conceptional age. This same pattern would be very abnormal in a
term patient and would be indicative of encephalopathy. Persistent dysmaturity is
associated with poor neurologic outcome. Transient dysmaturity may be due to non-
neurologic causes and is not necessarily associated with brain damage.
Visual analysis of neonatal EEG allows for detection of only great discrepancies, but this
is usually sufficient for routine interpretation. Quantitative analysis is possible but seldom
needed and not in routine use.
Background abnormalities
Excessive slow activity is difficult to discern, since neonates have prominent delta
activity already. Some infants with brain damage may have widespread delta, however.
The slow background is present in wake and sleep states and reacts poorly to exogenous
stimuli. This pattern is differentiated from normal delta activity by its widespread
distribution and lack of reactivity. Normal delta is prominent anteriorally and attenuated
by exogenous stimuli.
Amplitude asymmetries are significant only if they approach 50% of more. The
asymmetry usually indicates focal cerebral damage in the region of suppressed voltage. A
common pitfall is misinterpretation of asymmetries due to extracranial hematomas or
fluid collections. Subdural hematomas may suppress activity from one or both sides.
The isoelectric EEG is a confirmatory test for brain death. Guidelines for determination
of brain death are presented in Chapter 7.
Epileptiform activity
Epileptiform activity in the neonate can look very different from epileptiform activity in
older children and adults. The epileptiform activity may be focal, multifocal, or
generalized. Immaturity of cerebral maturation usually does not allow for generalization
of epileptiform activity.
Focal discharges occur usually in central region, more often on the right than the left. The
discharges may occur singly or in trains at 5-10/sec. Focal epileptiform activity is
differentiated from normal frontal sharp transients and multifocal sharp transients by
consistent lateralization. Also, normal sharp transients never occur in trains. The focal
discharges occasionally have a smooth contour and could be confused with an alpha or
theta rhythm. Sustained rhythmic activity is never normal in neonates of any conceptional
age, however. The rhythm must be differentiated from the fast component of delta
brushes by the absence of an underlying slow wave and the longer duration of the
epileptiform discharge than the fast component of a delta brush.
Focal discharges are usually associated with focal clonic seizures. The location of the
focus may not necessarily correlate well with the clinical seizure activity. The prognosis
for favorable neurologic outcome is good, since focal discharges in neonates do not
necessarily indicate a focal structural lesion.
Most focal sharp waves are surface negative. Surface-positive waves are seen in some
neonates with intracerebral hemorrhage. If the sharp wave is followed by a slow wave,
the hemorrhage is most likely subarachnoid. If the sharp wave is not followed by a slow
component, the hemorrhage may still be subarachnoid, but is more likely to be
ECN 3/e - 106
Rarely, neonates may manifest seizures without any perceptible alteration in background.
The generator of epileptiform activity is probably subcortical, and the discharges are not
projected to the surface. These infants usually have severe brain damage, explaining the
lack of rostral projection of the activity.
ECN 3/e - 107
• No pupillary reflexes
• No corneal reflexes
• No response to auditory or visual stimuli
• No response to “Doll’s head maneuver”
• No response to ice water caloric testing
• No respiratory effort with apnea testing.
The clinician must ensure that the absence of responsiveness is not due to drug
intoxication, metabolic disturbance, or neuromuscular blockade. Therefore, the following
parameters should be assured:
The period of observation can be shortened if there is a confirmatory test. These tests
include the following:
• EEG
• Brainstem auditory evoked potential
• Radionucleotide blood flow study
• Angiogram
Recently, transcranial doppler (TCD) has also been studied as a confirmatory test for BD.
Since BD is a complex legal issue and the President’s Commission did not specifically
mention TCD, this technique should not be used until the clinician can be assured that its
use for the determination of BD is part of accepted medical practice.
ECN 3/e - 109
If a confirmatory test is performed and is consistent with BD, the period of observation
can be reduced from 12 to 6 hours if the cause is not anoxia. The period of observation
can be reduced from 24 to 12 hours if the cause is anoxia.
• Do not declare a patient who is under the age of 7 days brain dead. Clinical and EEG
criteria are not established for this early period.
• If the patient is between 7 days and 2 months of age, perform two examination and
two EEGs 48 hours apart to determine BD.
• If the patient is between 2 months and 1 year of age, perform two examinations and
two EEGs 24 hours apart.
• If the patient is older than 1 year of age, perform two examinations 12 hours apart
without a confirmatory test. This observation period can be 6 hours if a single EEG is
done.
Despite these recommendations, most pediatricians do not feel comfortable with the
diagnosis of BD without a confirmatory test.
Table 7-2: Montage for determination of brain death in children. EEG for brain death
Channel Montage
1 Fp1-C3 Technical standards
2 C3-O1 include the following
3 Fp2-C4 recommendations:
4 C4-O2
5 Fp1-T3
6 T3-O1
7 Fp2-T4
8 T4-O2
ECN 3/e - 110
• Use a minimum of eight scalp electrodes covering all brain regions. This is usually a
reduced version of the 10-20 Electrode Placement System. The following electrodes
are recommended as a minimum: Fp1, Fp2, C3, C4, O1, O2, T3, T4.
• Use interelectrode distances of at least 10 cm. This allows for better detection of low-
amplitude EEG activity. A minimal montage would be that shown in table 7-2.
• Use interelectrode impedances that are no greater than 10 kohms but no less than 100
ohms. Too low an impedance occurs with electrode smear. The amplitude of recorded
electrocerebral activity will be excessively low if the impedance is low.
• Use a sensitivity of 2 µV/mm during most of the recording.
• Use a LFF setting of 1 Hz and a HFF setting of 30 Hz. [ck]
• Use ECG monitoring and other physiologic monitoring if necessary. Monitoring of
chest-wall motion may be needed if there is apparent slow activity in the record
which might be respiratory.
• Record reactivity of the EEG to auditory, visual, and tactile stimuli.
• Use a recording time of at least 30 minutes, most of which must be relatively artifact-
free recording.
• Test the integrity of the system by touching the electrodes to evoke a high-amplitude
artifact. This ensures that a flat background is not due to technical factors.
• Telephone transmission EEG cannot be used to support the diagnosis of BD.
• Recording should be made by a qualified technologist.
BD studies in children are performed in the same manner as BD studies in adults. More
physiologic monitoring is often required in children’s studies, however. Because of small
body size, respiratory movement artifact is relatively greater, and a chest-wall sensor is
desirable. An ECG channel is desirable for adult studies but is even more important for
BD studies in children. At high sensitivities, ECG artifact can be the predominant
potential in the record.
EEG monitoring
EEG is not always abnormal in patients with epilepsy. Therefore, long-term monitoring is
often needed to help with the diagnosis. Just as long-term cardiac monitoring is helpful
for evaluation of possible arrhythmia, long-term EEG monitoring can be invaluable. If a
patient has episodes which are possibly seizures but routine EEG is normal, EEG
monitoring should be considered. Most fellowship-trained neurophysiologists have
ECN 3/e - 111
adequate training and experience, but physicians who are not trained should not attempt
this interpretation. There are many potential pitfalls with these techniques.
There are two basic techniques: inpatient EEG monitoring and ambulatory monitoring.
Although the latter is fine, and we have done this for years, we have largely abandoned
this technique in favor of inpatient EEG monitoring. There is no substitute to being able
to see the event on videotape or streaming media as we see the EEG. Nevertheless, both
techniques offer invaluable information when making the important distinction between
seizure and pseudoseizure.
Monitoring laboratory
Dedicated EEG monitoring units are mainly the province of comprehensive epilepsy
centers, where patients are monitored for days in preparation for surgery. These units are
expensive but highly effective.
Most large hospitals have a sleep lab, and for a relatively small price, can be outfitted for
EEG monitoring. In some cases, this only entails adding software to the existing sleep-lab
equipment.
Portable EEG monitoring units are very helpful for patient in ICU, where there is concern
as to whether electrical discharges are controlled when the patient may not manifest signs
of obvious epileptiform activity. This equipment usually does not allow for recording of
the appearance of the patient, but since the patient is observed by the nursing staff, a
description of the event is typically reliable.
Electrodes and montages are the same as those used for routine EEG. In most
circumstances, a complete set of electrodes is placed. Digital acquisition systems allow
recording of each channel so that the montage and gain can be selected and changed
during the reading by the physician.
Collodion is used rather than paste for these long-term recordings. If the recording
exceeds 24 hours, the electrodes must be checked and gel re-applied to keep electrode
impedances sufficiently low.
ECN 3/e - 112
Performance
Many patients with pseudoseizures have multiple episodes per day, so that multiple days
of recording are not necessary. We usually perform a daytime 6-hour recording if routine
EEG is negative. If events are captured and can be interpreted, then further study is not
needed. If not, then 24-hour recordings are needed. We often take patients off of their
anticonvulsants for these long-term recordings, which makes both epileptic seizures and
pseudoseizures more likely.
Interpretation
Approach to interpretation
The interpreting physician will typically review all of the regions of interest. Seizure
discharges are reviewed along with segments of EEG before and after the discharge.
Clinical episodes will also be reviewed, using side-by-side comparison of the EEG and
video recording. Also, and regions of abnormalities identified by the technician will be
reviewed by the physician.
Seizure discharges
Seizure discharges may be subtle and easily overlooked on prolonged recordings where
there are multiple regions of interest. More commonly, the discharges are obvious;
generalized or focal spike-wave complexes can appear before and during clinical seizure
activity.
ECN 3/e - 113
Difficulty in interpretation may occur when muscle artifact obscures the recording during
the ictal event. This can be at least partly clarified by reducing the gain which can
eliminate some of the obscuration by artifact. Changing some filter settings can
accomplish some of this, as well. Unfortunately, changing filter settings can also alter the
frequency response of the system, altering spike as well as artifact configuration.
Pseudoseizures
Most of the studies performed in most EEG monitoring labs are for suspected
pseudoseizures. Any experienced neurologist will admit that history is often unable to
accurately able to differentiate epileptic seizures from pseudoseizures. Even direct
observation can be misleading. This difficulty is amplified by the fact that many patients
can have both epileptic seizures and pseudoseizures.
Quantitative EEG
Quantitative EEG has been a research tool for decades, but only recently has been
available to the practicing clinician. However, the clinical applications are limited, so that
QEEG is not part of routine practice. On first glance, digital signal acquisition lends itself
to quantitative analysis, possible making EEG interpretation more objective, less
dependent on the experience and bias of the interpreting physician. However, the
calculations reduce the amount of data which is considered in interpretation of the
information, thereby reducing some of the diagnostic power of the study. Also, the
calculated findings may not be significant for the patient. As has been discussed
elsewhere, amplitude asymmetries of up to 50% are normal for the waking posterior
dominant rhythm. This magnitude of difference would stand out o digital frequency
analysis yet be unimportant for the clinical interpretation of the record.
QEEG should be used as a tool which supplements rather than replaces conventional
visual EEG analysis. There is no substitute for the analytical ability of natures only neural
networks; QEEG findings should be interpreted with perspective as to their mathematical
origins and limitations.
Methods
Digital analysis begins with analog-to-digital conversion, as previously described. The
information from each channel is converted independently then stored and manipulated.
Montages are created by comparing the signal voltages from one channel with one or
ECN 3/e - 114
more others. For example, a longitudinal bipolar montage might be created by first
subtracting the signal from F3 from the signal from Fp1, etc. Usually, the native channel
data are stores along with a guide to the interpretive algorithms, so that these can be re-
selected at a later time. Filtering is performed with a broad frequency response.
Spike detection
Spike detection the most helpful feature of digital signal analysis. Spike detection
software performs frequency and amplitude analysis on the EEG record to identify
epochs which may contain epileptiform spikes. The software tends to err on the side of
detection, so most highlighted events are not true spikes, artifacts and physiologic activity
predominates. Nevertheless, the spike detection software allow the amount of EEG to be
reviewed to be reduced considerably. The neurophysiologists should review the EEG
before and after the spike to see if there is a state change, potential for artifacts, and
potentially ictal events.
Power spectral analysis is usually used in EEG for giving a visual impression of
frequency content. This can be particularly helpful for determination of encephalopathy
or sleep state. This an be used for intraoperative monitoring, especially. A relative
increase in slow activity or suppression of all activity over one hemisphere during
surgery is worrisome for infarction or some other insult to the underlying cortex.
Brain mapping
Brain mapping is the display of frequency data topographically. Brain mapping is most
helpful for detecting small asymmetries which would suggest a structural lesion. In this
regard, digital analysis is more sensitive than visual analysis for detecting these subtle
differences. Part of the limitation of brain mapping is that much of the data displayed is
interpolated, calculated from few actual recording points. This gives a nice appearance,
but may introduce data which did not exist. Minor differences may appear dramatic on
the mapping yet be clinically unimportant.
Brain mapping is used mainly for patients with seizures and patients with dementia. In
seizures, brain mapping can aid in the identification of areas of increased epileptiform
activity indicative of a focus. In dementia, quantitative EEG can increase the sensitivity
of routine EEG to detect mild slowing suggestive of an organic dementia rather than
pseudodementia.
Cerebrovascular disease has been studied in QEEG because changes in EEG are
immediate whereas imaging abnormalities may not be evident for days. Although this is
academically interesting, the EEG information does not currently have a clinical use in
routine care of acute stroke or TIA. Perhaps, in the future, EEG will help determine
whether a patient requires thrombolytics, since the EEG may be different in persistently
ischemic vs reperfused brain. However, MRI is more likely to be used in the stroke
centers.
ECN 3/e - 116
• Motor NCS
• Sensory NCS
Other nerve conduction studies are less commonly used, and include:
• F-wave study
• H-reflex
• Paired stimulation
• Repetitive stimulation at high and low rates
• Blink reflex
Needle EMG testing is performed in virtually all studies. The options for EMG are:
Other studies include the sympathetic skin response, and are included with special tests in
a separate discussion.
NCS and EMG are used for a wide variety of indications, and an individual approach is
usually needed. Some of the most common indications for study are:
Some of the most common diagnoses reached after NCS and EMG are:
ECN 3/e - 118
• Peripheral neuropathy
• Carpal tunnel syndrome
• Ulnar neuropathy
• Myopathy
Neuromuscular physiology
Normal neuromuscular function
We tend to think of the motor and sensory systems as separate, from motor and sensory
peripheral axons to the motor and sensory cortical regions of the brain. The anatomical
and physiological separation of these systems is greatly reduced with ascension in the
nervous system, and ascension in the evolutionary line. Yet, for this chapter, the motor
and sensory systems are tested separately.
Motor function
Descending input to the motoneurons in the spinal cord depolarize the dendrites. This
depolarization is conducted to the axon hillock where there is opening of voltage-
dependent sodium channels. When there has been sufficient depolarization to establish an
action potential, the efferent potential is conducted down the motoneuron axon to the
neuromuscular junction. The action potential depolarizes the nerve terminal which then
causes release of neurotransmitter into the junction. Binding of acetylcholine to receptors
on post-junctional muscle membrane produces depolarization of the muscle membrane.
When the depolarization is sufficient to generate an action potential, the potential is
conducted throughout the muscle fiber. This depolarization is causes release of
sequestered calcium, thereby facilitating muscle contraction. The contraction is
terminated when calcium is taken up by the sarcoplasmic reticulum and recycled for
another contraction.
Activation of one motoneuron results in activation of every muscle fiber in that motor
unit. One motoneuron action potential results in one muscle fiber action potential for a
wide range of firing frequencies. There is normally no spontaneous muscle fiber or
motoneuron discharge, meaning without descending or reflexive activation.
Muscle fibers associated with muscle spindles are intrafusal and muscle fibers
responsible for the power of contraction are extrafusal. These are innervated by separate
motor axons. The muscle spindles provide feedback to ensure that the extrafusal fibers
generate power for sufficient shortening. If the muscle does not shorten sufficiently with
the contraction, the frequency of discharge of active units is increased and inactive units
are recruited.
ECN 3/e - 119
Motor axons are of different size, with the axons for the fast-twitch (type II) motor units
being larger than the slow-twitch (type I) muscle fibers. This difference in motor axon
diameter is mirrored by a difference in diameter of the muscle fibers, with the fast-twitch
fibers often being larger in diameter than the slow-twitch fibers. This size relationship is
not exact, and differs between muscles, and between species. Physiological recruitment is
size dependent, with low-effort contractions generally recruiting smaller-diameter muscle
fibers. Again, this is variable, and large-diameter muscle fibers can be recruited first in
certain circumstance.
Sensory function
Electrical stimulation of the sensory nerves produces activation of potentially all of the
afferent axons, regardless of sensory modality. However, the large-diameter afferents
have the lowest threshold, whereas the small-myelinated and unmyelinated axons have
the highest threshold to electrical stimulation. Therefore, the submaximal stimulation of
EP studies typically does not activate these small fibers. Sensory NCS use maximal
stimulation, so the entire spectrum of sensory afferents is stimulated. However, the fast-
conducting large-diameter myelinated afferents contribute most to the SNAP, so the
sensory NCS tends to measure the conduction of the fastest fibers
The physiologic effect of each of these is summarized below and should be considered
when individual disorders are discussed.
Nerve cell body dysfunction is often due to neuronal degeneration, and the most
important disorders are ALS and SMA. We constantly lose neurons throughout our lives,
but in these and other neuronal disorders, cellular death is accelerated. The neuronal
ECN 3/e - 120
membranes become leaky so that there is influx of ions which depolarize the neurons.
This depolarization may occasionally reach threshold, producing spontaneous action
potentials in the neurons. In ALS, this is manifest as fasciculations.
The depolarization of the neuronal cell body results in activation of the voltage-
dependent sodium channels, but these channels are also time-dependent – they become
inactive after a short open time. The channels cannot open again until the membrane
potential is re-established, so if this does not happen, they cannot be activated again.
Therefore, the motoneurons become electrically inactive, contributing to the weakness.
The influx of ions, especially calcium, into the neurons results in activation or enzymes
including proteases and phospholipases which essentially digest the neurons from within.
This causes neuronal death, so the denervated muscle fibers try to find innervation from
surrounding surviving motor axons. Therefore, there are many more muscle fibers
innervated by a single motor axon, which is manifest on EMG as giant potentials. The
new nerve connections are not as fast or as consistent as the ones which developed in
early life, so there is a variety of times from motor axon activation to muscle fiber
activation which is manifest on EMG as polyphasic potentials.
Denervated muscle fibers develop their own fluctuations in membrane potential which
can occasionally reach threshold. The single muscle fiber action potentials are the basis
for fibrillation potentials and positive sharp waves. The difference between these two
patterns is in geometry of action potential generation and recording, so there is not a
pathological difference between these patterns.
NCS with motoneuron degenerations show reduced amplitude of the CMAP with little
change in NCV, since the conduction of the fastest fibers is little affected. Sensory NCS
is normal.
Axonal degeneration is the most common type of peripheral neuropathy, and there are a
multitude of causes. The common feature is degeneration of the distal portion of the axon
with denervation of the innervated muscle. This produces EMG changes which are
similar to those described for neuronal degeneration, including polyphasic potentials,
fibrillations, and positive sharp waves. While fasciculations can occur with axonal
degenerations, this is not as common as with neuronal degenerations.
NCS with axonal denervation is characterized by reduced amplitude of the CMAP and
SNAP, although the normal ranges of amplitude are so wide that amplitude, alone, may
not be sufficient to detect an abnormality unless there is prominent axonal drop-out.
NCVs are often normal but may be slightly slowed due to secondary demyelination.
Axonal damage results in partial unraveling of the myelin sheath. If an axon is still
functioning yet sick, the sheath may be dysfunctional enough to slow conductions.
Damage to the myelin sheath of peripheral nerve produces prominent slowing of the
propagation of action potentials. Normal axonal conduction is fast because the myelin
sheath increases the impedance of the axonal membrane. Depolarization of the axon
results in electrotonic conduction of the depolarization to the next node, or gap in the
myelin sheath. The axon membrane at the node is capable of generating an action
potential, but the membrane between the nodes, underneath the myelin sheath, is not.
Electrotonic conduction is virtually instantaneous, compared with action potential
propagation, so with myelinated axons, the action potential essentially skips from node to
node. This conduction is much faster than action propagation down an axon. The myelin
sheath greatly reduces the decay in electrotonic conduction [diagram] which would
normally occur, facilitating the fast conduction.
Muscle dysfunction
ECN 3/e - 122
The most important myopathies are muscular dystrophies and inflammatory myopathies
(polymyositis and dermatomyositis) Myopathies produce damage to the muscle fiber
membrane. Changes in conductance to ions results in influx of sodium which can
occasionally reach threshold. This produces fibrillation potentials and positive sharp
waves which are indistinguishable from those of neuronal degenerations. However, the
motor unit potentials differ. There is not reinnervation so there are no long-duration
polyphasic potentials or giant potentials. Rather, each motor unit has fewer function
muscle fiber potentials giving small motor unit action potentials. The conduction in the
muscle fibers is dispersed, so that the muscle fiber potentials is somewhat separated. This
gives rise to polyphasic potentials, but the duration is short since the change in
conduction is less than with axonal damage. Therefore, myopathic motor unit potentials
are small-amplitude and short duration. This is called brief small-amplitude polyphasic
potentials (BSAPPs).
Equipment
Equipment used in electrophysiological assessment of nerve and muscle is similar in
overall design and function. Most equipment is essentially a computer with interface card
for signal acquisition and controlling a stimulator. The stimulator is a waveform
generator controlled by a controller module in the computer. The acquisition equipment
consists of amplifiers, wide-band filters, and is external to the computer, itself. The
output from the acquisition equipment is then fed to an analog-to-digital conversion
module of the computer.
normal in appearance and low in amplitude, a dropout of axons is usually the culprit.
This is seen in pathological processes which primarily affect axons. If the waveform is
much broader and has several phases, then demyelination is usually to blame, causing
dispersion and conduction block.
Electrodes: For surface stimulation, the stimulator electrodes are usually stainless
steel and mounted 2-3 cm apart on a small hand-held two-pronged probe. By convention
the cathode (negative pole) under which negative charges collect is black in color. The
anode (positive pole) is denoted by a red color. In most situations the cathodes of the
stimulating and recording electrodes are adjacent on the nerve being studied and the
anodes face away from each other. Checking that the electrodes are set up “black-to-
black” aids in troubleshooting if problems arise. Certain studies require needle electrodes
for stimulation. In this case, the cathode consists of a small needle inserted into the skin
near the nerve and the anode may be either a surface electrode or another needle. Much
less stimulator current is required when needle electrodes are used because the impedance
of subcutaneous tissue is lower than that of skin, and the stimulating electrodes are much
closer to the nerve under study.
Electrode
position: The
stimulating electrodes
are placed on the skin
overlying the nerve at
two or more sites along
the course of the nerve.
The recording
electrodes are placed
over the belly of the
muscle, with the active
electrode over the
Figure 8-1: Motor nerve conduction study.
midbelly of the muscle,
A: Diagram of the right forearm and representation of the electrode as close to the estimated
positions for median nerve conduction study. Photos of exact electrode endplate site as
positions are included on the ccompanion CD. possible.
B: Sample recording of median motor compound motor acction
potentials (CMAPs).
Stimulus
characteristics:
Stimulation of the nerve being studied is accomplished using a brief burst of direct
current. Stimulators are of two types: constant voltage and constant current. In our
laboratory, constant current stimulation seems to provide the most consistent responses.
Constant current stimulators vary the voltage of stimulation to compensate for changing
skin impedance, providing a consistent current to the nerve being stimulated, while
ECN 3/e - 124
constant voltage units vary the current to obtain consistent voltage to the nerve.
Regardless of type, the amount of current applied to the tissue is never more than about
100 mA, and rarely exceeds 500-600 volts. Stimulus duration is usually in the range of
50-300 milliseconds for most studies. One exception is the H-reflex study which
typically requires a much longer stimulus duration on the order of 500 to 1000
milliseconds for best results.
Stimulus artifact can often be troublesome, especially in the case of sensory NCS. It can
arise from numerous causes. Helpful in reducing stimulation artifact are: cleansing
recording and stimulation skin sites with alcohol, placing the ground electrode between
stimulation and recording sites, ascertaining that good electrical contact exists at all
contact points, reducing stimulation intensity and/or duration, and increasing the distance
between stimulation and recording sites when possible.
Procedure: After all of the electrodes are in place, the instrument is set to deliver
repetitive stimuli, usually at 1 Hz. The stimulus voltage is initially set to zero then
gradually increased with successive stimuli. A CMAP appears that grows larger with the
increasing stimulus voltage. Eventually, further increases in voltage do not cause any
change in CMAP amplitude. A stable response is assured if the voltage is 25% greater
than the voltage needed to produce the highest amplitude CMAP. Once a good recording
is made, the trace is stored for later analysis and the stimulating electrode moved
proximally to a second stimulus site. Most nerves are stimulated in two sites for motor
nerve conductions, but some are stimulated in at least three locations along the course of
the nerve. It is not necessary to gradually increase the stimulus intensity for the
subsequent electrode positions since the patient is now used to the concept of electrical
stimulation and would probably like to minimize the number of stimuli delivered.
Table 8-1: Stimulus and recording parameters for motor and sensory NCS.
Parameter Motor NCS Sensory NCS F wave H reflex
Gain 2 mV/division 20 µV/division 200 µV/division 200 µV/division
Time base 2 msec/division 1 msec/division 10 msec/division 10 msec/division
LFF 10 Hz 10 Hz 10 Hz 10 Hz
HFF 32 kHz 32 kHz 32 kHz 32 kHz
Stimulus duration 0.2 msec 0.1 msec 0.2 msec 0.2 msec
Interpretation
Measurements: the following measurements are made from the CMAP produces
by stimulation at each site:
The distance between stimulus sites is measured, and the nerve conduction
velocity calculated according to the following simple formula:
Dist
NCV =
( PL − DL)
Where Dist =distance, PL = proximal latency, and DL = distal latency. The final results
are expressed as meters per second or m/sec.
Types of abnormalities: The most common types of abnormalities in motor NCV are:
There are defined norms for nerve conduction velocities, and these are presented in Table
8-2.
Slow motor NCV: Slowing of the NCV below the normal range indicates a defect in the
myelin sheath such that the axons do not conduct as fast as they normally should. This
could be from neuropathy, nerve entrapment, cooling, and other causes. Comparing one
NCS to others helps with this classification.
Increased distal latency: Increased DL of the CMAP from the most distal site indicates
slowing of conduction in the most distal portion of the motor nerve. This can be due to
peripheral neuropathy although in this case there would usually be slowing of conduction
is more proximal segments of the peripheral nerve.
Relative slowing of motor NCV by comparing segments: The absolute motor NCVs
may be normal but there may be a discrepancy between the velocities of segments of the
nerves. The best example of this is ulnar entrapment across the elbow, where a difference
in 10 mm/sec in motor NCV is significant even with normal absolute NCVs.
Decreased amplitude or altered waveform of the CMAP: These abnormalities are not
given the degree of importance that velocity changes are. Nevertheless, marked reduction
in amplitude suggests axonal drop-out. Severe damage to the myelin sheath can disperse
ECN 3/e - 126
the waveform so that the amplitude is less, but in genera, amplitude correlates with
axonal dysfunction more than myelin dysfunction.
Electrodes: The
electrodes are essentially
the same as those used
for motor NCV. In
addition spring-like
stainless steel rings can
be used for the fingers
for stimuation or
recording. The cathode is
placed facing the
stimulator. Certain near-
Figure 8-2: Sensory nerve conduction study. nerve technique studies
A: Diagram of the right forearm for stimulation and recording of the use a small needle
median nerve sensory nerve action potential (SNAP).
B: Sample recording of the SNAP. inserted in the skin near
the nerve as the recording
surface. In sensory NCS the active and reference electrodes are placed along the nerve
about 3-5 cm apart. Placing the ground between stimulus and recording sites can
effectively reduce artifact. It is important to ask the patient to relax the limb being
studied. This simple maneuver can especially increase the quality of sensory NCS.
Electrode location: In order for the nerve response to be purely sensory, either the
stimulating or the recording electrode has to be on a purely sensory portion of the nerve.
For motor NCV, the recording is obtained from the muscle, so only motor fibers
contribute to the response. With sensory studies, stimulation of a mixed nerve will
activate both motor and sensory fibers so that recording has to be over a distal sensory
branch of the nerve for a SNAP to be recorded. Alternatively, the stimulation could be
reversed, so that the stimulation is on the distal sensory branch with recording from the
ECN 3/e - 127
proximal nerve trunk. Both ways are used. When the stimulus is distal, so that the nerves
are conducting in there usual afferent direction, this is antegrade conduction, when the
connections are reversed, this is retrograde conduction.
Averaging: Most modern equipment has the capacity to average the results of several
stimulations. This is most helpful in sensory NCS where the signal-to-noise ratio is much
lower and the typical working voltages are one or two orders of magnitude less than those
seen in motor NCS. Use of the averager is quite useful in bringing out elusive low-
amplitude sensory nerve action potentials.
Interpretation
Dist
NCV =
LO
where Dist = distance between the stimulating and recording electrodes and LO is the
latency to the onset of the sensory potential. Unlike motor NCS, both the onset and peak
latencies can be used for interpretation, and most of us use peak latency at a defined
distance as a measurement rather then calculating a conduction velocity.
Low amplitude: The amplitude of the response is normally quite low, such that
averaging is often needed in order to get a reproducible and measurable response.
Therefore, little significance is given to reduced amplitude as long as the latencies and
conduction velocities are normal.
ECN 3/e - 128
F-wave study
The F-wave tests conduction of motor axons proximal to the stimulation site. When
motor nerves are stimulated for routine NCS, the stimulation creates action potentials that
travel not only orthodromically toward the muscle but also antidromically toward the
motoneuron. The antidromic potential reaches the some and depolarizes the dendrites.
Depolarization is then conducted electrotonically back to the axon hillock, which is now
repolarized. A new action potential is created and transmitted back to the muscle. The
action potential activates the motor end-plate, causing action potentials in muscle fibers.
This late response is called the F-wave.
ECN 3/e - 129
For the F-wave study, the electrodes are placed as for the motor NCS. The stimulating
electrode is over the nerve either proximally or distally. However, the stimulating
electrodes are turned around so that the cathode (negative pole) is toward the spine.
H-reflex
The H-reflex is the
electrophysiologic counterpart of
the Achilles tendon reflex. The H-
reflex is elicited by stimulation of
the tibial nerve in the popliteal
Figure 8-4: H-reflex study. fossa while the patient is lying
H-reflex study elicited by stimulation of the tibial nerve prone. The feet may be supported
and recording from the soleus. The stimulus intensity is by a pillow to give a slight bend at
gradually increased. The H-reflex is elicited first by the knee. Recording is made using
activation of the muscle afferents. The afferents synapse surface electrodes over the soleus or
in the spinal cord and response returns in the
motoneurons With further increase in stimulus intensity medial gastrocnemius. The stimulus
intensity is gradually increased and
the H-reflex appears at a lower intensity of stimulation than the CMAP. As the stimulus
intensity is further increased, the CMAP appears and the H-reflex disappears. The muscle
afferents are larger and have a lower threshold for electrical activation than the alpha
motoneurons, hence they are activated at a lower threshold of stimulation.
Normal H-reflex latency is normally 35 msec or less, and interside differences should not
exceed 1.4 msec. H-reflex amplitude differs widely between patients and should not be
used as a criterion, if the potential is visible and of normal latency.
ECN 3/e - 130
Blink reflex
The blink reflex can be used to evaluate
patients with lesions of the facial nerve
or of the brainstem. However,
neuroimaging is the preferred method to
evaluate the brainstem, so the blink
reflex is used predominantly for
evaluation of cranial nerves 5 and 7.
• Prolonged direct response with otherwise normal latencies indicate a lesion of the
facial nerve, such as Bell’s palsy.
• Prolonged R1 indicates a lesion in the reflex pathway from trigeminal nerve to facial
nerve. Facial nerve lesions also prolong R1, but are distinguished from brainstem or
trigeminal nerve lesions because the direct response is prolonged, as well. If the
contralateral R2 is normal, then the afferent limb in the trigeminal nerve is normal,
and a brainstem lesion is likely.
• If the latency of the direct response is normal but the latencies of R1, ipsilateral R2,
and contralateral R2 are prolonged, a lesion of the trigeminal nerve is likely, but
brainstem lesions cannot be excluded.
Electromyography Basics
Methods
Electrodes: In needle electromyography (EMG) the recording surface is typically the tip
of a small sharp needle inserted through the skin into the muscle under study. In the case
of monopolar EMG needles, the reference and ground electrodes are small surface
electrodes of the type used in NCS. In the case of concentric needles, the reference
electrode is the barrel of the needle and the active electrode is part of the needle tip
electrically distinct from the barrel. Monopolar electrodes are used predominantly,
nowadays.
Procedures: The surface electrodes are placed first then the needle inserted into the
muscle to be studied. Then the amplifier is turned on to begin the recording. Recordings
should be made in the following categories:
• Rest
• Insertion
• Single motor unit activation
• Maximal contraction
Insertional activity Movement of the needle in an otherwise Brief volley of action potentials.
relaxed muscle
Motor unit Needle is not moved while patient makes A few motor unit action potentials,
potentials slight contraction. biphasic or triphasic, short duration
Single motor unit activation: After assessment of rest and insertion activity, the patient
is asked to make a slight contraction of the studied muscle. This will evoke motor unit
potentials (MUP) which are generated by each motor axon activating many muscle fibers.
The motor unit may innervate 100-2000 muscle fibers, depending on the muscle tested.
Muscles chosen for study: The number and location of muscles chosen for study
depends upon the clinical question. In the case of myopathy and proximal weakness,
proximal muscles in at least two limbs are tested. The choice of muscles for EMG study
should always be tailored to the specific clinical situation of each individual patient.
Extreme care should be exercised when performing needle EMG in patents currently
being treated with anticoagulant medication. In our laboratory we seldom perform EMG
when the INR is greater than 2.0 or the PTT greater than 45.
Interpretation
Normal and abnormal responses
Resting activity: Normal resting muscle is electrically silent. There will be small
amplitude of noise, but no sign of muscle fiber or motor unit activation. Many disorders
of peripheral nerve and muscle produce instability of the muscle fiber membrane. This
results is episodic degeneration of action potentials which then is manifest as single
muscle fiber action potentials. These are fibrillation potentials and positive sharp waves,
discussed above. They have an amplitude of about 50 µV and are less than 3-5 ms in
duration. Occasional MUPs may be seen, and when they occur in a regular rate, are
probably due to incomplete relaxation and if otherwise normal in appearance, should not
ECN 3/e - 134
Needle insertion may evoke myotonic discharges. These are repetitive muscle fiber action
potentials. The pattern of repetition is characteristic and has a waxing and waning quality.
This suggests to some a “dive bomber” quality on the audio monitor, but is more like a
motorcycle accelerating and decelerating. The waveform itself is initially a negative slow
onset followed by a faster positive component, similar to the positive sharp wave. A
number of conditions can cause myotonic activity, discussed below.
Motor unit potentials: The physiology of normal and abnormal motor units was
discussed earlier in this chapter. Normal motor units should be biphasic or triphasic with
a duration of 10 msec or less. Longer duration units suggest motor unit reorganization of
reinnervation. Polyphasic potentials also suggest reinnervation. If the polyphasic
potentials are normal or large amplitude and long-duration (>10 msec) this suggests
reinnervation. If the polyphasic potentials are of short duration and especially of low
amplitude, this suggests myopathy.
Maximal contraction: The recruitment pattern with maximal contraction is seldom the
only sign of abnormality. Decreased recruitment means that there are fewer functioning
motor units, so there is not the complete obscuration of the baseline which should occur.
Early recruitment means that the baseline is obscured at low levels of effort. This is
because there are ineffectual contraction of sick muscle fibers so that more motor units
and muscle fibers are recruited.
Myopathy
In myopathy the muscle fibers are smaller and have less membrane area. Parts of the
muscle fiber are incapable of supporting an action potential. Neuromuscular junctions
may be unstable or dysfunctional because of inflammation or degeneration of the
membrane at or near the NMJ. As a result of these processes, myopathic motor units are
lower in amplitude, shorter in duration, and tend to be polyphasic and unstable. The
characteristic sound produced on EMG is very scratchy. It sounds a little like sandpaper.
Motor unit recruitment is abnormally rapid, with full interference pattern produced with
relatively little effort.
Neuropathy
ECN 3/e - 137
rates of 50-60 HZ, requiring 50-100 stimulations. While the information gained from
such testing is quite useful, the test is uncomfortable and is no longer offered in many
laboratories. Similar information can be gained from examining pre-exercise and post-
exercise RNS tests. A decrement in amplitude between the first and last stimulation of
greater than 10% is indicative of abnormal NMJ function.
History should focus on the pattern of symptoms and their development and progression
in time. Is there any numbness, weakness or pain? Where is it and how severe, what are
exacerbating and alleviating factors? Examination should include testing of sensory and
motor function as well as reflexes. Armed with the clinical history and examination
findings enables us to tailor the testing to the individual patient’s needs. The worst
complication of electrophysiological testing is failure to get the needed information, or
worse, to muddy the water with misinformation.
ECN 3/e - 139
Mononeuropathy is extremely common in clinical practice. While almost any nerve can
suffer entrapment or injury at one time or another, some are seen more commonly than
others and we will confine our discussion to those most likely to be encountered in the
typical electrophysiology laboratory.
Median nerve
Median motor NCS is performed by placing the active recording electrode on the
midbelly of the APB over the thumb. The reference is placed 2 cm distal. The cathode for
distal stimulation is 7 cm proximal to the active recording electrode near the wrist crease.
The proximal site is over the median nerve proximal to the antecubital fossa. In both
instances, the anode is placed 2 cm proximal to the cathode.
Median sensory NCS is performed by recording from one of the fingers, usually the
index, with ring electrodes while stimulating 13 cm proximally, above the wrist crease.
The stimulating and recording electrodes can be reversed, and the results may differ
slightly depending on the direction of conduction.
The most common entrapment neuropathy seen in our laboratory is Carpal Tunnel
Syndrome, (CTS) which is a distal median neuropathy within the carpal tunnel at the
wrist. Clinical diagnosis is usually straightforward, and the characteristic
electrophysiological findings no less so. The most common finding is slowing of distal
sensory NCV and/or prolonged median nerve terminal latency, in the absence of
abnormalities in other tested nerves. It is commonly bilateral and usually worse in the
dominant hand. Denervation in the ABP muscle is sometimes seen in severe cases and
warrants more aggressive treatment typically consisting of surgical decompression of the
nerve in the carpal tunnel.
ECN 3/e - 140
Ulnar nerve
Ulnar sensory NCS is performed by recording from digit 5 using ring electrodes and
stimulating the ulnar nerve in the distal forearm 11 cm proximal to the active recording
electrode.
Radial nerve
Motor NCS is performed by recording from the extensor indicis. The active recording
electrode is placed over the belly of the muscle and the reference is placed 2 cm distal.
The distal stimulation site is in the forearm, between the ECU and extensor digitorum
ECN 3/e - 141
Peroneal nerve
Tibial nerve
The tibial nerve innervates the medial and lateral gastrocnemius and soleus muscles and
supplies sensation to those portions of the sole and dorsolateral foot that are not served by
sural and superficial peroneal nerves. The tibial nerve also innervates most of the intrinsic
muscles of the foot via the medial and lateral plantar nerves. Tibial sensory NCVs are
rarely performed and will not be discussed. Motor NCVs are performed by recording
from the belly of the abductor hallicus muscle on the medial aspect of the foot. Distal
stimulation is delivered to the nerve as it passes behind the medial epicondyle, and
proximal stimulation is delivered in the popliteal fossa.
Sural nerve
Sciatic nerve
Sciatic neuropathy is seen less commonly but often can masquerade as peroneal
neuropathy. In most sciatic nerve injuries the peroneal division bears the brunt of the
injury with relative sparing of the posterior tibial division. The reason for this is unclear.
EMG revealing denervation in the distribution of both divisions of the sciatic nerve is
helpful in establishing this diagnosis. NCS are not performed on the sciatic nerve, per se,
but rather on the individual divisions: peroneal and tibial.
Femoral nerve
ECN 3/e - 143
The femoral nerve is supplied by roots from the lumbar plexus, and L2-L4 make up most
of the innervation. NCS of the femoral nerve can be technically challenging, so most of
the diagnosis rests on EMG. Denervation in the quadriceps suggests a femoral
neuropathy.
Chapter 9: Approach to
clinical questions
Overview
No test is worthwhile in the absence of clinical information. All electrophysiological
tests are merely extensions of the physical examination. Failure to focus on the relevant
history and examination is an open invitation to misdiagnosis. A good bat in the hands of
a poor batter will not likely result in an acceptable hitting record, and a bat stuck in the
ground is even less likely. Do a brief history and a focused examination at some point in
the testing process and definitely prior to any EMG.
History should focus on the pattern of symptoms and their development and progression
in time. Is there any numbness, weakness or pain? Where is it and how severe, what are
exacerbating and alleviating factors? Examination should include testing of sensory and
motor function as well as reflexes. Armed with the clinical history and examination
findings enables us to tailor the testing to the individual patient’s needs. The worst
complication of electrophysiological testing is failure to get the needed information, or
worse, to muddy the water with misinformation. The first task of neuromuscular
diagnosis is to correctly classify the disorder.
Radial neuropathy Radial motor NCS. Radial sensory NCS Extensor digitorum communis. Triceps
Sciatic neuropathy Tibial motor NCS. Peroneal motor NCS. Tibialis anterior. Medial gastrocnemius. Vastus medialis.
Sural sensory NCS. Tibial F wave. Peroneal Gluteus medius
F wave
Spinal stenosis see Lumbar radiculopathy see Lumbar radiculopathy
Tarsal tunnel Medial plantar motor NCS Lateral plantar Abductor hallucis
syndrome motor NCS. Tibial motor NCS. Sural
sensory NCS
Thoracic outlet Median motor NCS. Median sensory NCS. Abductor pollicis brevis. First dorsal interosseus. Extensor
syndrome Ulnar motor NCS. Ulnar F wave. Ulnar digitorum communis
sensory NCS
Ulnar neuropathy Ulnar motor NCS across and below elbow.. First dorsal interosseus. Abductor digiti minimi
Ulnar sensory NCS.
ECN 3/e - 146
The most common finding among patients with generalized neuromuscular disorders is
axonal neuropathy. This has a huge differential diagnosis and a distinct diagnosis is not
established in all patients.
Important neuropathies
The most common neuropathies encountered in routine clinical practice are:
• Neuronal degenerations
• Polyneuropathies
• Mononeuropathies
ECN 3/e - 148
Neuronal degeneration
Amyotrophic lateral sclerosis
ALS is the most common neuronal degeneration affecting the PNS. There is
degeneration of upper and lower motoneurons without involvement of sensory systems.
Motor and sensory NCVs are normal, although the CMAP my be of reduced amplitude.
EMG shows widespread denervation. Examination for motoneuron disease including
ALS should include examination of at least three muscles in different nerve distributions
in each of three limbs. The head may be substituted for one limb, and tongue examination
is quite easy. Electrical signs of acute and chronic denervation are seen, but acute signs
may be subtle, depending on the stage of the disease. Fasciculations are common in ALS,
but by themselves are not signs of denervation, since benign fasciculations may occur.
Poliomyelitis
Poliomyelitis is caused by a neurotropic enterovirus that destroys anterior horn
cells. Polio is rare today, and we have personally only identified one case, related to
vaccine. Motor NCVs are normal or near normal. EMG shows chronic denervation in
survivors. Since polio is focal or multifocal, the EMG findings are most prominent in
clinically-affected muscles. When polio affects patients at a young age, the degree of
polyphasia is less but the MUPs are larger, indicating better reinnervation of denervated
muscle fibers. Signs of motor unit reorganization are often seen in muscles which were
clinically unaffected.
ECN 3/e - 149
Post-polio syndrome has received tremendous attention of the lay and medical
communities. Patients complain of a variety of symptoms, including increasing weakness
and pain. EMG shows signs of chronic denervation with high-amplitude polyphasic
MUPs, with some occasional fibrillation potentials. The clinician must determine whether
there are any signs of new disease responsible for the symptoms. The neurophysiologist
can, in most instances, document the prior denervation on the basis of EMG alone.
Prominent signs of active denervation should suggest new axonal damage. Repetitive
stimulation and single fiber EMG are of little benefit for evaluation of weakness in
patients with post-polio syndrome since polio will produce abnormalities.
Polyneuropathies
Table 10-2: Important polyneuropathies.
Polyneuropathy Clinical features NCS and EMG
Diabetic sensorimotor neuropathy Numbness and weakness distally. NCS may be normal but more
Often with burning few. commonly mildly slow. Chronic
denervation
Diabetic mononeuropathy Common mononeuropathies are Multifocal slowing and
multiplex ulnar, peroneal, femoral. denervation, especially ulnar,
peroneal, femoral nerves.
Hereditary motor sensory Distal weakness and numbness. Slow NCVs, increased DLs, F-
neuropathy Type I Decreased reflexes. AD waves delayed or absent.
(Charcot-Marie-Tooth) inheritance.
Hereditary motor-sensory Distal weakness with less NCVs normal or mildly slowed.
neuropathy Type II suppression of DTRs than in type Motor DLs increased. CMAP and
(Neuronal form of CMT) I. AD inheritance. SNAP may be decreased.
Hereditary motor-sensory AR inheritance. Distal weakness Slow NCVs. Denervation is less
neuropathy Type III and numbness. prominent.
(Dejerine-Sottas)
Hereditary motor sensory CNS signs as well as peripheral Slow NCVs. Denervation is less
neuropathy Type IV neuropathy. Upper motoneuron prominent.
(Refsum’s disease) signs.
CMAP amplitude between distal and proximal stimulation sites. Many neuropathies such
as those seen in diabetes mellitus and uremia cause varying degrees of axonal loss and
demyelination.
Diabetic neuropathy
DM causes 4 basic types of neuropathy:
• Small-fiber polyneuropathy
• Large-fiber polyneuropathy
• Autonomic neuropathy
• Mononeuropathy and mononeuropathy multiplex
The large-fiber neuropathy involves both motor and sensory axons, with particular loss of
motor fibers, and vibration, joint perception, and two-point discrimination.
NCV is slowed in most patients with diabetic neuropathy even though the demyelination
is often secondary to axonal degeneration. The EMG usually shows denervation in
clinically-affected muscles. Patients with isolated mononeuropathies or radiculopathies
superimposed on a polyneuropathy will often have fibrillation potentials and neuropathic
MUPs in clinically affected muscles.
EMG evidence of denervation is unusual but may develop in severe cases. Reduced
recruitment and conduction block are seen early in the course.
When demyelination is severe, permanent axonal loss may occur in distal muscles. EMG
and NCV usually return to normal, but conduction slowing may persist in severe cases.
There is an axonal form of AIDP, which might seem like an oxymoron. However, the
clinical presentation can be very similar, though some reflexes may be preserved.
Some clinicians follow serial NCVs to assess response of IDP to treatment. This is likely
to be more routine with future practice.
• Vincristine
• Cisplatin
• Lead
• Ethanol
There are no specific findings on NCS and EMG to differentiate between toxic axonal
neuropathies and other causes of axonal neuropathy.
Hereditary neuropathies
Hereditary motor-sensory neuropathy type I
ECN 3/e - 152
HMSN-II is the neuronal form of CMT. It is clinically similar to HMSN-I, however, the
NCS and EMG findings are very different. Motor and sensory NCVs are either normal or
mildly slow, indicating relative preservation of myelin. Motor distal latency may be
increased, and CMAP and SNAP amplitudes may be reduced. The EMG shows signs of
acute and chronic denervation, including prominent fibrillation potentials, fasciculations,
and high-amplitude long-duration polyphasic MUPs
A focal form of HMSN-III has been described. Only one nerve is affected. NCVs in other
nerves are normal. The EMG is normal except in the distribution of the affected nerve,
where there is prominent acute and chronic denervation.
HMSN-IV is Refsum’s disease, with both central and peripheral nervous system
involvement. In the peripheral nervous system, demyelination causes slowed NCVs.
Degeneration of anterior horn cells causes EMG features of acute and chronic
denervation.
Friedreich’s ataxia
Mononeuropathies
Entrapment Neuropathies
Entrapment neuropathy is extremely common in clinical practice. While almost any
nerve can suffer entrapment at one time or another, some are seen more commonly than
others and we will confine our discussion to those most likely to be encountered in the
typical electrophysiology laboratory.
Median nerve
The most common entrapment neuropathy seen in our laboratory is Carpal Tunnel
Syndrome, (CTS) which is a distal median neuropathy within the carpal tunnel at the
ECN 3/e - 154
Routine median motor and sensory conduction study may miss patients with clinically
significant CTS. While at least 90% of patients with CTS will have some abnormality,
this leaves many patients without objective proof of their lesion. To improve the
detection of CTS, palmar stimulation and comparative studies are often performed.
Palmar stimulation consists of stimulating the median nerve at 1 cm increments on either
side of the palmar crease while recording the median SNAP. Patients with CTS will
usually have slowed conduction across one or more of these small segments. When
performing incremental stimulation, it is important to not use excessive voltages or
stimulus durations, otherwise the site of nerve activation may be far from the stimulation
site, thereby compromising the effect of the technique.
Pronator teres syndrome is due to compression of the median nerve as it passes through
the pronator teres muscle in the proximal forearm. Median motor NCV through the
forearm is slow, but the distal latency is normal. Sensory NCV is normal because the
segment tested is distal to the pronator teres.
EMG shows active and chronic denervation of the abductor pollicis brevis, flexor
digitorum superficialis, and median innervated portion of the flexor digitorum profundus.
The pronator teres is not denervated because it is innervated proximal to the site of
compression. This is an important feature that distinguished the pronator teres syndrome
from compression of the median nerve by the ligament of Struthers.
ECN 3/e - 155
The ligament of Struthers is a fibrous band above the medial epicondyle. The clinical
syndrome of median nerve compression by the ligament of Struthers resembles the
pronator teres syndrome. Most median-innervated muscles of the forearm are weak. The
differentiating feature is involvement of the pronator teres. In pronator teres syndrome,
the pronator teres is tender but strength is normal, and there is no denervation on EMG.
When median nerve compression occurs at the ligament of Struthers, the pronator teres is
weak and shows acute and chronic denervation on EMG.
The anterior interosseus nerve is a branch of the median nerve that innervates several
muscles in the forearm, including:
This syndrome is caused by injury to the nerve as it leaves the main trunk of the median
nerve. NCVs are usually normal. Stimulation of the anterior interosseus nerve at the
elbow may reveal increased distal latency of the CMAP recorded from the pronator
quadratus. EMG shows denervation in the flexor pollicis longus, median-innervated
portion of the flexor digitorum profundus, and pronator quadratus.
Ulnar nerve
The ulnar nerve is susceptible to injury at the wrist, in the forearm, at the elbow, and
above the elbow. In practice, the most common cause of ulnar neuropathy which we see
is in diabetes, where ulnar nerve damage is prominent and stands out more severely than
other components of neuropathy.
The ulnar nerve passes from the forearm into the hand through Guyon’s canal. Ulnar
compression at the wrist is analogous to CTS. Unlike compression of the ulnar nerve at
the elbow, the ulnar-innervated flexors in the forearm are unaffected, and the sensory loss
is confined to the ulnar side of the hand, sparing the forearm.
Ulnar NCVs are slowed across the wrist and the DL is increased when recording from the
ADM. EMG shows denervation in the ADM and 1st-DI, but a normal pattern in the FDP
(ulnar side) and FCU.
The ulnar nerve is particularly susceptible to injury near the elbow. Compression is
exacerbated by occupation or habit placing pressure on the ulnar groove or proximal
forearm, obesity, propensity to acute elbow flexion, and polyneuropaties which
predispose to pressure palsies.
NCS shows slowing of motor NCV across the elbow and slowing of distal sensory
conduction. Occasionally, the lesion can be primarily axonal in nature and very little
slowing is seen but the EMG shows denervation in the ulnar-innervated hand intrinsic
muscles. Denervation is not seen in the flexor carpi ulnaris because the nerve branch
innervating this muscle leaves the ulnar nerve proximal to the elbow.
The palmar branch of the ulnar nerve innervates ADM and provides sensation overlying
digit 5. Therefore, lesion of this produces normal routine NCS including ulnar motor DL.
However, DL to the 1st-DI is prolonged and denervation is seen in this muscle. The
superficial sensory branch of the ulnar nerve may be damaged, especially in bicycle
riders.
Radial nerve
The most common sites of damage to the radial nerve are at the spiral groove in the arm
and in the forearm as the nerve penetrates the supinator muscle. Injury can also occur
with dislocation of the elbow and distally from compression by handcuffs.
Spiral groove
Radial nerve entrapment or damage in the spiral groove in the humerus is not rare and
typically follows a night of overindulgence. This is the so-called Saturday Night Palsy.
Wrist drop and finger drop are seen along with loss of sensation in the distribution of the
radial sensory nerve. Abnormalities in radial sensory NCS are seen and in motor NCS.
The amount of denervation in radial muscles seen on EMG is variable. This entity is
easily distinguished from C7/C8 radiculopathy by demonstrating normal function in
median and ulnar-innervated muscles.
As the radial nerve enters the forearm, it divides into the posterior interosseus nerve and a
superficial sensory branch. The posterior interosseus nerve supplies the finger and wrist
extensors. Injuries to the nerve cause weakness without sensory loss.
Radial NCV is slowed through the involved segment. EMG shows denervation in the
wrist and finger extensors with sparing of the extensor carpi radialis longus and
supinator. Both muscles are innervated by branches that arise proximal to the lesion.
Proximal muscles are more severely affected than distal muscles. Intrinsic muscles of the
hand are spared. The most prominent denervation is in the deltoid, biceps, supraspinatus,
and infraspinatus. The serratus anterior and rhomboids are spared because their
innervation is proximal to the lesion.
The upper plexus is most comonly damaged by stretch, such as when the shoulder is
forced down by impact or by pull downward of the arm. The upper eplexus herves are
stretched and temporarily cease to conduct action potentials. If the stretch is sufficiently
severe, there may be breakage of the axons, which can result in long-term impairment.
Radiation plexopathy is another important cause of upper plexus damage. There is less
surrounding tissue around the upper plexus than the lower plexus, so radiation is more
likely to have long-term deleterious effects.
Brachial plexitis
NCVs may be normal or slow through the plexus. Severe lesions may cause slowing of
distal median and ulnar conduction. The EMG may sow denervation not only in weak
muscles but also in muscles that seen clinically unaffected. Sufficient time must elapse
before the EMG is abnormal, however. In practice, EMG shows denervation at about the
time that the patient has developed significant weakness. Therefore, if there is pain but no
weakness, the initial EMG may be normal and a later study may be more revealing.
Peroneal neuropathy
Peroneal nerve entrapment at the fibular head producing foot drop is the most common
entrapment syndrome of the lower extremity seen in our laboratory. This is commonly
seen in trauma, but in our practice, we see this more commonly as a complication of bed
rest in hospitalized patients.
ECN 3/e - 158
Slowing of peroneal NCV at the fibular head is the classic finding. Denervation in
tibialis anterior and peroneus muscles is variably seen. Care must be exercised to
exclude the possibility of L5 radiculopathy and sciatic neuropathy. In the case of the
former, denervation and weakness in posterior tibialis and hamstring muscles is seen in
addition to findings in muscles supplied by the common peroneal nerve. In sciatic
neuropathy, an abnormal EMG in the biceps femoris short head is sufficient to establish a
more proximal localization of injury.
Tibial neuropathy
Tibial nerve entrapment behind the medial malleolus is called tarsal tunnel
syndrome. This condition is relatively uncommon, but frequently considered by
podiatrists and orthopedic surgeons ordering NCS & EMG. The motor DL of the medial
plantar nerve, lateral plantar nerve, or both are often increased. The medial plantar nerve
is tested by stimulation of the tibial nerve with recording from the abductor hallucis. The
lateral plantar nerve is tested by recording from the abductor digiti quinti. Sensory
conduction through the tarsal tunnel is performed by stimulation of the medial or lateral
plantar nerves and recording from the tibial nerve behind the medial malleolus. Because
of the tremendous effect of foot temperature on nerve conduction, especially sensory,
subtle differences in latency should be interpreted with caution, and the absence of a
SNAP on the affected side should support the clinical diagnosis. Motor distal latencies
are less dependent on temperature than sensory conduction. Differences in left-right and
medial-lateral conductions are evidence of tarsal tunnel syndrome.
Some neurophysiologists use incremental stimulation through the tarsal tunnel, similar to
that performed for carpal tunnel syndrome. This probably does not add anything to
careful measurement of medial and lateral plantar motor DL and sensory latency,
however.
Sciatic neuropathy
Sciatic neuropathy is seen less commonly but often can masquerade as peroneal
neuropathy. In most sciatic nerve injuries the peroneal division bears the brunt of the
injury with relative sparing of the posterior tibial division. The reason for this is unclear.
EMG revealing denervation in the distribution of both divisions of the sciatic nerve is
helpful in establishing this diagnosis.
Pirifomis syndrome
the biceps femoris; in peroneal entrapment, this muscle is unaffected since innervation
arises from the peroneal division proximal to the knee.
Sciatic stretch
The sciatic nerve can be damaged by stretch of the sciatic nerve which can occur
especially in the lithotomy position – abduction, and slight flexion of the legs. This
position stretches the sciatic nerve as it passes through the sciatic notch into the leg. The
patient awakens from anesthesia wit weakness of sciatic-innervated muscles. This
syndrome can also be seen in patient who have sudden forward flexion at the waist.
NCVs of the tibial and sural nerves are usually normal. F-wave latency may be increased.
EMG is normal or shows only a decreased number of MUPs immediately after the injury;
electrodiagnostic study is often performed soon after the injury, before electrical signs of
denervation develop.
Radiculopathy
The search for radiculopathy is a daily task for the electromyographer. Radiculopathy
results from irritation of the nerve root as it exits the spinal cord. Most irritative lesions
are extradural but the inflammatory process can occasionally extend somewhat into the
subarachnoid space, giving rise to abnormalities in the CSF. Needle EMG assessment is
the cornerstone of diagnosis in these disorders. In choosing muscles for study, the
clinical history and neurological examination is of paramount importance. Muscles
within the suspected abnormal myotome should be selected, along with muscles from
adjacent myotomes. As an example, for a suspected C7 radiculopathy, one could study
the bicep, tricep, pronator teres, anconeus, first dorsal interosseus, and abductor pollicis
brevis. Abnormalities found exclusively in C7-innervated muscles traveling in different
peripheral nerves can clearly reveal the process as a radiculopathy as opposed to a
mononeuropathy. When the clinical picture is vague and the examination normal, a
representative sampling of muscles innervated by several levels is appropriate. An
example for the upper extremity would be as follows: deltoid, supraspinatus, (C5); bicep,
brachialis, (C6); tricep, pronator teres, (C7); extensor indicis proprius, first dorsal
interosseous, abductor pollicis brevis, (C8). Routine nerve conduction studies should
always precede electromyography to eliminate confusion. If there is denervation in all
tested C8 muscles for example, NCS can help determine whether this is due to a distal
generalized neuropathy. Radiculopathy can often produce abnormalities of mild slowing
and decreased amplitudes in the motor NCS. with relatively normal sensory NCS. This
pattern coupled with characteristic abnormalities in the EMG is helpful in determining
the diagnosis of radiculopathy. Table 10-4 described commonly seen radiculopathies and
their respective EMG findings.
Mononeuropathy multiplex
Diabetes mellitus and polyarteritis nodosa are the most common causes of multiple
mononeuropathies in the US. Leprosy is a common cause worldwide.
Polyarteritis nodosa
Affected nerves show slowed or blocked conduction at the site of arteritis. EMG shows
acute and chronic denervation in affected areas. Complete denervation cases fibrillation
potentials and positive sharp waves and absence of MUPs with attempted voluntary
effort.
Leprosy
Leprosy is probably one of the most common causes of neuropathy worldwide. The
neuropathy is caused by either primary nerve infiltration or by infarction of the vasa
nervorum.
ECN 3/e - 161
NCS shows blocked or slowed conduction. EMG shows active and chronic denervation
in muscles innervated by affected nerves If the neuropathy is predominately due to
cutaneous vasculitis, only the distal sensory branches may be involved.
ECN 3/e - 162
• dystrophies
• inflammatory myopathies
• metabolic myopathies
Motor and sensory conduction velocities are normal in most disorders, although selected
metabolic disorders affect peripheral nerves in addition to muscles. CMAP amplitude
may be reduced because muscle fibers fail to be activated. EMG is essential for
diagnosis. Insertion may elicit complex repetitive discharges. At rest, there are fibrillation
potentials and positive sharp waves. Motor unit potentials (MUPs) are reduced in
amplitude and brief in duration. With increasing effort, units are recruited earlier than
normal because of reduced tension output of the muscle fibers, termed early recruitment.
Muscular dystrophies
Dystrophic disorders – Duchenne and Becker’s muscular dystrophy
Duchenne and Becker’s muscular dystrophy are characterized by normal NCVs and
myopathic findings on EMG. EMG shows fibrillation potentials, complex repetitive
discharges, and early recruitment. Fibrillation potentials are not as prominent as with
inflammatory myopathies and with denervation. In later stages of the disease, muscle is
replaced by fat and connective tissue, and insertional activity is reduced or absent.
Limb-girdle dystrophy
ECN 3/e - 163
Myotonic dystrophy
Myotonic dystrophy is characterized by myotonic on EMG. Myotonia is the repetitive
discharge of muscle fibers with an initially high-frequency that gradually declines. This
produces a “dive bomber” sound in the audio monitor. Myopathic MUPs may also be
seen. Occasional neuropathic features may include slow motor conduction and a reduced
number of functioning motor units.
Inflammatory myopathies
Electrophysiologic changes in inflammatory myopathies are all the same. NCVs are
normal, although CMAP amplitude may be reduced. EMG findings include fibrillation
potentials and positive sharp waves, myopathic MUPs, and complex repetitive
discharges. Fibrillation potentials are more prevalent in inflammatory myopathies than in
muscular dystrophies. Abnormalities are more prominent in clinically weak muscles. The
EMG is normal in approximately 10% of patients with polymyositis. This may be due to
sampling error or to periods of relative inactivity during the course of the disease.
Metabolic myopathies
Mitochondrial disorders may present as neuropathy or myopathy. Therefore, NCV and
EMG should both be done when a mitochondrial myopathy is suspected. Tibial motor
NCV and sural sensory NCV are a sufficient screen. If the sural SNAP is absent with a
normal tibial motor NCV, a sensory NCV in the arm or of the superficial peroneal
sensory NCV should be performed.
• Cushing’s syndrome
• Addison’s disease
• Thyrotoxicosis
• Hypothyroidism (rarely)
• Hyperparathyroidism (rarely)
NCVs are normal except in hypothyroidism, in which case they may be slowed. EMG
may show minor myopathic features in all of these disorders.
Steroid myopathy (Cushing’s syndrome) is usually a clinical diagnosis. The EMG usually
does not show myopathic features. The NCVs are also normal, although CMAP
amplitude may be reduced. Biopsy can help to differentiate inflammatory myopathy from
steroid myopathy in patients being treated for inflammatory myopathy, but is seldom
necessary.
Table 11-3: Metabolic myopathies
Disorder Clinical findings NCS and EMG
Mitochondrial myopathies Weakness with other Myopathic MUPs, but findings
manifestations, e.g. ptosis, may be subtle or absent.
ophthalmoplegia, cardiac May have neurogenic appearance,
abnormalities. as well.
Mild increase of CK
concentration.
Myotonia congenita Myotonia with or without Myotonia
cramps. RNS may show decrement with
Onset in youth or young adults. increasing stimulation frequency.
AD, AR, or sporadic Some patients have myopathic
MUPs.
Hypokalemic periodic paralysis Episodic weakness provoked by Normal between episodes.
rest after large carbohydrate Reduced MUPs during attack.
meal. Reduced CMAP amplitude.
Hyperkalemic periodic paralysis Episodic weakness provoked by Between attacks: possible
rest after exercise or cold. myopathic patterns.
During attack: myotonia,
increased insertion, reduced
MUPs to volition to stimulation.
Carnitine palmityl transferase (CPT) deficiency usually manifests normal NCV and EMG
findings. Patients with carnitine deficiency have myopathic findings with small-
amplitude polyphasic MUPs. Fibrillation potentials are seen but are rare.
• Tetanus
• Stiff-man syndrome
• Schwartz-Jampel syndrome
• Neuromyotonia (Isaac’s syndrome)
Tetanus
Tetanus is characterized by involuntary discharge of motor units. The toxin works at the
spinal level, blocking postsynaptic inhibition, thereby increasing the excitability of the
motoneurons. The EMG shows repetitive MUPs that are abolished by peripheral nerve or
neuromuscular block. The discharges are attenuated during sleep and with general or
spinal anesthesia.
Stiff-man syndrome
Stiff-man syndrome is not really a disorder of muscle; it is due to excessive motoneuron
activation. The reason for the enhanced discharge is unknown. Excessive motor unit
activation results in involuntary muscle contraction involving predominantly proximal
muscles. Affected muscles show normal MUPs with coactivaion of agonists and
antagonists. Discharges are attenuated by sleep, general anesthesia, benzodiazepines,
peripheral nerve block, or neuromuscular block.
Schwartz-Jampal syndrome
Schwartz-Jampel syndrome is characterized by multiple congenital anomalies in
association with increased muscle fiber activity. The defect is probably at the nerve
terminal. Clinically, the muscle activity looks similar to that of myotonia, but on EMG
the discharges have the appearance of complex repetitive discharges, lacking the
frequency modulation of true myotonia. The discharges are abolished by neuromuscular
block but not by nerve block.
Neuromyotonia
Neuromyotonia is a term used to differentiate muscle fiber activity of nerve terminal
origin from myotonia which is of muscle membrane origin. Neuromyotonia is repetitive
activity of single muscle fibers rather than of complete motor units. The muscle fibers
discharge repetitively at frequencies that are initially high and gradually decline. This is
similar to true myotonia, but the discharges have an invariant decline in frequency rather
than a waxing and waning frequency. Also, these discharges are apparent at rest, whereas
myotonia is evoked primarily by needle insertion. The amplitude of MUPs may be
reduced because of the loss of functioning muscle fibers due to continuous activity. The
defect is probably in the terminal motor axon. Therefore, the discharges are abolished by
neuromuscular block but not by peripheral nerve block, spinal block, or general
anesthesia.
ECN 3/e - 167
NCVs are normal. EMG shows myotonia. Repetitive stimulation produces a decremental
response which is greater with high frequencies of stimulation.
Paramyotonia congenita
Paramyotonia congenita (of Eulenberg) produces weakness with myotonia which is
exacerbated by repetitive activation. Potassium is increased, as in hyperkalemia periodic
paralysis. Cold can evoked the weakness with myotonia, but with progressive chilling the
myotonia disappears and the muscles are flaccid. Repeated contractions exacerbate the
myotonia, in contrast to myotonia congenita in which repeated contractions alleviate the
myotonia.
NCVs are normal. Repetitive stimulation often results in a decremental response. EMG
shows myotonia with needle insertion or percussion.
Periodic paralysis
The periodic paralyses are a family of disorders all characterized by abnormal loss of
excitability of the muscle fiber membrane. The common thread is the symptom of
episodic muscle paralysis, however, the clinical presentations are quite diverse.
Chapter 12:
Neuromuscular junction
defects
Overview
Three principal conditions comprise 99% of all cases of NMJ transmission defects seen
in the average EMG laboratory: myasthenia gravis (MG), Lambert-Eaton Myasthenic
Syndrome (LEMS) and botulism. MG is a post-synaptic defect and LEMS and botulism
are presynaptic defects with distinct pathophysiology. Defects in NMJ function cause
weakness of affected muscles with no sensory loss. In addition to abnormal results in
RNS testing, increased jitter in SFEMG is seen with all NMJ disorders.
Table 12-1: Disorders of neuromuscular transmission.
Disorder Clinical features NCS and EMG
Myasthenia gravis Weakness that worsens with Normal CMAP.
activity. RNS – decremental response with
Ptosis and diplopia. low rates.
Abnormal SFEMG with
increased jitter and blocking.
Lambert-Eaton myasthenic Generalized or proximal RNS – decremental response at
syndrome weakness. low rates, incremental response at
Dry mouth, impotence, and/or high rates.
other signs of autonomic Facilitation with exercise.
dysfunction.
Botulism GI distress followed by weakness Low CMAP amplitude, increases
with bulbar involvement. with exercise.
RNS – little decrement at low
rates, increment at high rates.
Repetitive stimulation and single-fiber EMG are performed on patients who are being
evaluated for the possibility of myasthenia gravis. Paired stimulation is performed mainly
ECN 3/e - 169
for botulism. LEMS is associated with abnormalities on both repetitive stimulation and
single-fiber EMG testing.
Table 12-1 summarized the clinical and neurophysiologic features of these disorders.
Myasthenia Gravis
Myasthenia gravis (MG) is a condition caused by production of IgG antibodies directed
against the acetylcholine receptor on the post-synaptic membrane. This binding
stimulated internalization and degradation of the receptor, therefore, there are fewer
receptors available for binding with acetylcholine. When an action potential depolarizes
the presynaptic membrane, the transmitter cannot activate sufficient receptors to evoke an
action potential in the muscle fiber. The sarcolemmal depolarization is not sufficient.
Botulism
Botulism is a condition caused by a specific neurotoxin elaborated by several strains of
C. botulinum. The toxin consists of two peptide chains. The long chain is responsible
for binding of the peptide to the cell surface and entry of the toxin. The short chain is the
active neurotoxin. It is a serine metalloprotease which cleaves certain peptides in the
presynaptic nerve terminal responsible for vesicle transport. This makes neurotransmitter
release impossible, and permanently inactivates the synapse. The toxin is incredibly
potent and it has been estimated that as little as four molecules of toxin are required to
inactivate a single neuromuscular junction. A half liter of pure toxin would be sufficient
to kill all people living on the earth.
Clinically, the syndrome of botulism is distinguished from the other NMJ disorders by
signs of autonomic involvement, chief among which is paralytic ileus. Nausea, vomiting,
constipation, and other GI symptoms are common early in the course.
Electrophysiologically, findings are similar to those seen in LEMS with the exception
that an incremental response at high rates of stimulation is not seen. EMG findings
indicative of denervation also develop within a few weeks, which also helps distinguish
this NMJ disorder from the others. NCS is normal except for reduced CMAP amplitude.
Successive stimuli result in further reduction in CMAP amplitude. Paired stimuli can be
helpful for diagnosis. At short intervals – less than 15 ms – the response to the second
pulse is greater than to that of the first. This is because the second impulse activates some
ECN 3/e - 171
terminals that were not activated by the first pulse due to summation of end-plate
potentials.
Neonatal myasthenia
A classification of neonatal myasthenic syndromes was presented by Misulis and
Fenichel (1989) based on pathophysiology. Diagnosis of these syndromes depends on
techniques not readily available in most laboratories, including:
Routine NCS and EMG are performed on all patients to look for neuropathies or
myopathies which could be confused with genetic myasthenia. Transitory neonatal
myasthenia occurs in children of myasthenic mothers. Repetitive stimulation at 3 Hz
produces a decremental response.
ECN 3/e - 172
Overview
Evoked potentials are the electrical responses to sensory stimulation. While there are
motor evoked potentials, they are not in routine clinical use and will not be discussed
here. The three clinically used EPs are brainstem auditory (BAEP), visual (VEP), and
somatosensory (SEP).
With the advent of MRI, the role of EPs in neurologic localization has waned, somewhat,
however, there is still some benefit to functional studies. One of the most persistent roles
of EPs has been in evaluation of patients for possible MS. Patients present with
paraparesis, hemiparesis, or visual loss, and may have MRI evidence of white-matter
lesions. EPs can show functional abnormalities in affected and clinically unaffected
regions. This documentation of multiple lesions supports the diagnosis of MS.
Visual loss usually triggers two types of evaluation - ophthalmological and MRI. Careful
ophthalmological evaluation can determine whether the visual loss is pre-chiasmatic or
post-chiasmatic. Funduscopic changes may suggest optic neuritis, increased intracranial
pressure, vascular abnormality, or primary retinal disorder. When the lesion is not felt to
be ocular, MRI is commonly done to look for structural lesion. When none is found, EPs
can document abnormality in optic nerve function which can suggest retrobulbar optic
neuritis.
These indications are for evaluation of clinical lesions without visualization on structural
imaging. EPs are also helpful for assessment of clinically silent lesions in suspected
multiple sclerosis. In this circumstance, the somatosensory EP is most valuable for
patients with optic neuritis, and visual EP is most valuable for patients who present with
paraparesis.
There are many generators of EPs, since responses are recorded from multiple sites along
the afferent projection pathways. The VEP is most likely due to charge movement
associated with conduction in projections from the lateral geniculate to the visual cortex,
Central SEP activity is due to thalamo-cortical projections, but impulses conducted in the
peripheral nerves and dorsal columns are recorded as well. The BAEP is recorded from
the nerve volley in the eighth cranial nerve and potentials generated by tracts and nuclei
in the brainstem. Specific locations of generators are discussed in the individual sections
on VEP, BAEP, and SEP.
The generators of EPs are of two basic types: nerve bundles and nuclei. Nerve fiber
bundles include both peripheral nerves and central tracts. The recorded potential is due to
the advancing front of the compound action potential. The vector of this potential is
determined by the direction of projection of the axons.
Potentials generated in nuclei are not easily described by vectors and axonal conduction.
Movement of charge in nuclei is a combination of axonal action potentials and charge
movement during synaptic transmission. Synapses are oriented in virtually all directions
on a cell's dendrites and soma, such that it is impossible to predict the ultimate vector of
positivity and negativity. Also, because of the complex orientation of the synapses, there
is no guarantee that the field will conform to a simple dipole. Therefore, hypotheses of
the sources of individual EP waves are developed on the basis of human pathology and
animal studies in addition to a knowledge of basic neuroanatomy.
The equipment used to record EPs is similar to that used for routine EEG and EMG
studies and should fulfill the guidelines for electrical safety outlined in Part I. In general,
modern and well-maintained machines meet the required limitations on allowable leakage
ECN 3/e - 175
current. Unsafe practices can endanger a patient even with the best equipment, however.
The greatest risk is with SEPs, since the stimulus is an electrical pulse. The patient must
be adequately grounded so that the path of current cannot traverse the heart or spinal
cord. Normally, the path of current is between the two electrodes of the stimulator,
however, current can flow from stimulating electrode to ground if one lead has poor skin
contact or high impedance. Minimal technical requirements for EP machines are
tabulated in Table 13-1.
Acquisition of signals
Stimulus depends non the EP modality, and is covered in the respective chapters. The
common thread the stimulus-response circuit for each of the modalities. There are central
projections for the sensory inputs, and the central responses to those inputs produces the
normal EP.
Averaging
EPs are of very small amplitude and in most instances cannot be seen without averaging.
The EP is superimposed on EEG activity which is unrelated to the stimulus. In addition,
muscle electrical activity and movement artifact contributes to the recorded potential.
Averaging brings out the EP by the assumption that most potentials not caused by the
stimulus occur in a random fashion and will not produce a potential of substantial
amplitude after averaging many trials. This assumption is generally true but has two
possible sources of error. First, the stimulus may cause a slight movement of the patient
that is sufficiently reproducible from trial to trial to be detectable in the average. An
experiences neurophysiologist can usually identify such abnormal waveforms. Second,
60 Hz interference can appear to be a high-amplitude sinusoidal wave on averaging. To
prevent this latter error, the stimulus rate should not be a harmonic of 60 Hz. The
ECN 3/e - 176
Artifact rejection
Analysis time
A time window is set during which the response is acquired. The sampling begins shortly
after the stimulus. A brief delay is used to reduce stimulus artifact affecting the input
amplifier. A high-voltage stimulus artifact may be conducted to the scalp electrodes, and
subsequently peak the input amplifier. The high potential can alter the amplification of
the amplifier for a brief time, a fraction of a second, but this may be time enough for the
response to be missed. This is most important for SEPs, but there can also be artifact
during BAEP from current movement in headphone leads. Pattern reversal VEPs have
effectively no stimulus artifact, though flash VEPs may produce potentials in anterior
(usually reference) leads; this can be minimized by good electrode contact.
Limiting the duration of the recording was more important on earlier machines with low
operating and storage capacities, but effectively is eliminated as a factor now.
Nevertheless, there is no use to recording long after the response has passed. Modern EP
equipment allowed for changing of sweep speed which in turn alters recording duration
and sampling interval.
Replications
Two replications of each waveform are recommended for each EP. Consistency of
waveforms is visualized if the traces are superimposed on the hard copy. Four
replications may be necessary when recording SEPs to provide convincing identification
of individual waves.
EP recordings are all bipolar. Up or down on the trace depends on which electrode is
active and which is reference. For the VEP, this is clear, but it is not so clear for SEPs
and BAEPs. For EEG, a negative potential delivered to the active electrode is shown as
ECN 3/e - 177
an upward deflection off the pen. Polarity conventions have recently become fairly
consistent for EPs, and in general, tend to show the potentials of clinical importance as
upward deflections. For SEPs and VEPs, traces are displayed so that a negative event at
the active electrode produces an upward deflection. In contrast, for BAEPs, a positive
event at the active electrode produces an upward deflection.
Normal data from Vanderbilt University Hospital are shown in the respective sections.
Testing equipment and environment differ between laboratories, however, and each lab
should set its own set of normative data, or at the very least, ensure that data obtained
from normal subjects fits within the established published norms. For published norms to
be used, not only must the data be verified in the lab, but the techniques of acquisition
must be identical.
There are maturational changes in EPs early in life, so the Guidelines recommend that
normative data be established for each week of the perinatal period, for each month of
infancy, and each decade, thereafter. At least 20 subjects from each age group should be
tested for each evoked response. Responses from the left and right sides of the same
subject cannot be considered to be two subjects. Such data are used to establish
normative data on interside differences in latency and amplitude.
Reports
Reports should be concise but thorough. A sample report is shown in Figure 13-1. It is
most helpful to put the data in tabular form. Highlighting abnormal values is also helpful.
The interpretation can have two sections. The first described what is abnormal, and the
second gives the implications of the abnormalities. Interpretation of the data in light of
the clinical history is essential, since many physicians ordering EPs are not experts in this
field.
ECN 3/e - 178
Hard copies of the waveforms should be kept with the patient's record in the laboratory.
Most modern equipment allows for selected waves to be printed on the report along with
the tabular data. This is helpful for other neurophysiologists, but is not off interest for
most clinicians.
ECN 3/e - 179
The VEP is the only evoked response that is visible without averaging. During routine
EEG, photic simulation is used to activate epileptiform discharges in patients suspected
of having seizures. The photic simulation is a bright flash delivered o subjects with heir
eyes closed. At flash frequencies of less than 5-7/sec, an evoked response is recorded
from the occipital leads. A driving response is recorded at faster frequencies. The VEP is
highly reproducible as long as the patients maintains fixation and has no change in visual
acuity.
Methods
Stimulus
The VEP stimulus may be:
• flash,
• full-field pattern reversal, or
• half-field pattern reversal
Flash is used in patients who cannot cooperate with the level of fixation required for
pattern reversal stimulation. The latencies of the flash-VEP are more variable than the
pattern reversal VEP, so the flash VEP can really only test continuity of the visual
pathways. Full-field pattern reversal is the usual stimulus for the VEP. Each eye is
examined individually, so the anterior visual pathways are evaluated especially well.
Half-field pattern-reversal simulation is used for localization of lesions behind the optic
chiasm. Although many laboratories still perform half-field testing, modern imaging
ECN 3/e - 180
Flash
The strobe light is similar to that used for photic stimulation during conventional EEG.
The strobe is placed in front of the patient's eyes, usually with the eyes closed. Plenty of
light passes through the lids to activate the retina.
Presence of the flash VEP indicates continuity of the pathways from the retina to the
lateral geniculate. flash VEPs have been recorded in the absence of a functioning cortex.
Therefore, flash stimuli are not used if reproducible waveforms can be obtained with
pattern-reversal stimuli.
Pattern-reversal
Check size affects the latency and amplitude of the VEP. Size is measures in minutes of
visual field arc where there are 60 minutes (60') per degree of arc. The maximal response
ECN 3/e - 181
is elicited by a check size between 15' and 60'. The wide range of check size is made
possible by differences in other variables, notably size of the stimulus field. With smaller
checks, the latency of the response is increased and the amplitude is decreased. The fovea
is stimulated better by smaller checks, and the peripheral is stimulated better by larger
checks. The recommended check size is 28' o 32' of visual field arc, which is a
compromise between the two extremes.
Table 14-1: VEP stimulus and recording parameters.
Parameter Value
Reversal rate 2/sec
Contrast > 50%
Check size 28-32 min or arc
Field size 8º of arc
Number of trials 100-200
Recommended electrodes
Midline occipital (MO) 5 cm above inion
Right occipital (RO) 5 cm right of MO
Left occipital (LO) 5 cm left of MO
Midline frontal (MF) 12 cm above nasion
Ear (A1) Left ear or mastoid
Recommended montages
Channel 1 LO – MF
Channel 2 MO – MF
Channel 3 RO – MF
Channel 4 MF – A1
Recording parameters
LFF 1.0 Hz (-3 dB)
HFF 100 Hz (-3 B)
Analysis time 250 msec
Measurements N75 latency from each eye
P100 latency from each eye
Interocular latency difference
Amplitude (baseline to P100 peak or N75 to P100
peaks)
Interocular amplitude ratio (larger/smaller)
Stimulus field size should be at least 8 degrees (8º) of the visual field arc, since
approximately 80% of the response is generated by the central 8º of vision. A smaller
field size has been recommended to increase sensitivity to subtle defects, but false
positive rate is unacceptable. Visual acuity is the limiting factor in the presence of
reduced stimulus field and reduced check size.
Reversal rate should be 2/sec, which is an interstimulus interval of 500 ms. Faster
reversal rates cause an increase in the latency of the major wave, P100. Rates faster than
5-7/sec produce the entrained driving response seen on routine EEG during photic
stimulation.
Luminance is standardized since low levels increase P100 latency and decrease
amplitude. Standard computer monitors produce sufficient luminance for routine VEP
testing, but this should be checked periodically. Note that the increasingly-popular LCD
monitors often have a lag in luminance after being switched on, and some have slow
ECN 3/e - 182
refresh rates, compared to CRT monitors. Pupillary diameter can affect effective retinal
illumination, so marked interside differences in pupillary diameter must be considered in
interpretation of the VEP. Although there are no specific recommendations for luminance
levels, it is recommended that luminance level remain constant between studies.
Contrast between the light and dark squares must be greater than 50%. In practice, the
contrast is much greater than this. Low contrast results in a delayed and lower amplitude
P100.
Fixation on the target has more of an effect on amplitude than latency. Poor fixation
results in reduction in amplitude, but as the response is detectable, there is much less
effect on latency. Some individuals are able to voluntarily reduce the P100 amplitude
sufficiently to make the P100 unidentifiable, partly through fixation.
Half-field stimulation is delivered to one eye at a time and one half-field, either right or
left. The pattern reversal technique is he same as that for full-field stimulation, with the
checkerboard on one side and the screen blocked out. Comparing the responses with
stimulation of the two half-fields, tests the visual pathways behind the optic chiasm.
ECN 3/e - 183
MRI and CT provide excellent visualization of the retrochiasmal visual pathways and are
superior to VEP for detection of lesions in these regions. Therefore, half-field stimulation
is not commonly used for evaluation of patients with suspected brain pathology.
The left and right half-fields are alternately stimulated with the stimulated area at least
one check-width from the fixation point. One eye is stimulated at a time. Montage is
shown in Table 14-2.
Recording
Electrode placement and montage
Recommended electrode placements and montage are presented in Table 14-1. these are
in keeping with the Guidelines (American Electroencephalographic Society, 1994), and
are used by most neurophysiology labs.
Response from the midline-occipital (MO) electrodes is used for most VEP
interpretation. The MO electrode is active and the midline-frontal (MF) electrode is the
reference. Electrodes on either side of MO are right occipital (RO) and left occipital
(LO); these can be helpful for interpretation of abnormal studies, but these responses are
not used if MO responses are normal. Some laboratories use a Cz-Oz derivation, which is
satisfactory, but labs should conform to the newer guidelines.
The LO and RO electrodes help with waveform identification, especially for half-field
stimulation. For routine pattern-reversal stimulation, placement of electrodes is dictated
by findings. If the waveforms are poor or unusual, use of lateral electrodes, use of more
anterior electrodes or both is recommended to look for an unusual potential distribution.
Recording parameters
Recording parameters are presented in table 14-1. Filter settings are standard, with a
bandwidth of 1 to 100 Hz. Analysis time is 250 ms for most machines. Default recording
parameters are set for EP machines, although they are changeable by the software.
Interpretation
Waveform identification
Inspection of the normal VEP reveals three identifiable waveforms: N75, P100, and
N145. The P100 is a positive potential at about 100 ms, and is the only one used for VEP
interpretation. The negative potentials at about 75 ms and 145 ms help with identification
of the P100, but are too variable and inconsistent for routine interpretation
ECN 3/e - 184
Variant waveforms
The two most common variant waveforms are the bifid pattern and inverted waveform.
Both are due to variation in anatomical orientation of the visual cortex and optic
radiations.
Bifid pattern: The P100 is split into two humps, which can be confusing. The
neurophysiologist should treat this like a camel with one or two humps. Measure the
hump from the peak of a single hump beast, or interpolate between the humps for a two
hump beast. This interpolation [ck erm usage] is shown in figure x-x. If the interpolated
P100 is normal, then he study is interpreted as normal, though comment of the variant
waveform should be made in the impression.
Occasional patients will have such a widely split P100 that is not amenable to
extrapolation. This may be due to defects in the projection to the upper and lower
segments of the calcarine cortex, possibly from visual field defects. Stimulation of only
the lower half of the visual field may improve the bifid waveform, but the latency is often
abnormal, anyway. There is controversy as to whether o not a bifid waveform should be
considered abnormal on its own. The most conservative approach is to simplify the
waveform by lower-half stimulation and localize the P100 by recording from Pz and Cz
in addition to midline occipital derivations. Using these techniques, the rest is abnormal if
he latency of the resultant waveform is increased.
ECN 3/e - 185
Abnormalities
The types of abnormalities include:
Clinical correlations
Optic neuritis
Optic neuritis typically increases the latency of the P100 of the PR-VEP. If the optic
neuritis is purely unilateral, then the increase is usually also unilateral. A prolonged
latency of the P100 from an asymptomatic eye suggests a previous subclinical episode of
optic neuritis. A sample VEP from a patient with optic neuritis is shown in Figure 14-3.
ECN 3/e - 186
Multiple sclerosis
Approximately 15% of patients with
optic neuritis will later develop other
findings leading to the diagnosis of
MS. In patients with optic neuritis,
SEP is frequently performed to look
for clinically silent lesions of the
spinal cord. VEP testing may be done
to document the visual defect in the
clinically affected eye and to evaluate
Figure 14-3: VEP in optic neuritis.
Left optic neuritis produces a delay in the VEP
the optic nerve of the unaffected eye.
produced by stimulation of that eye. Abnormal VEP latency is present in
about 40% of patients with multiple
sclerosis who do not have a history of optic neuritis. Virtually all patients with a history
of optic neuritis have either an absolute increase in the latency of the P100 from the
affected eye or an abnormal interside difference in latency, if the absolute latencies are
normal.
Tumors
Tumors arising in the region of the anterior optic pathways produce compression of optic
neuritis and/or chiasm. This results in visual field defects that affect each eye differently.
The VEP is almost always abnormal, but the correlation between visual acuity and degree
of VEP abnormality is poor. Typical abnormalities are alterations in absolute or interside
latencies of the P100 and changes in wave morphology and amplitude. Latency changes
are more reliable than morphology or amplitude changes.
Tumors affecting the posterior visual pathways are less likely to affect the VEP. Full-
field pattern-reversal VEP testing is usually normal in patients with dense hemianopia.
The use of electrodes lateral to MO may reveal an amplitude asymmetry with the higher
amplitude ipsilateral to the side of the lesion, but amplitude asymmetries may be present
in normal individuals. half-field stimulation reveals abnormalities in some individuals,
however, the sensitivity and specificity are not good enough to justify using this
technique to screen for posterior lesions. Imaging techniques should be used.
Pseudotumor cerebri
Patients with pseudotumor cerebri have increased intracranial pressure which is no due to
a mass lesion, venous thrombosis, or other structural defect. If untreated, the increased
pressure can produce visual loss. If treatment is effective, the visual loss can improve,
however permanent deficits result if there has been long-standing increased pressure.
Most patients with pseudotumor cerebri have normal VEPs. A few patients are described
with abnormalities in association with incipient visual loss, however, VEPs should not be
ECN 3/e - 187
A normal flash-VEP indicates continuity of the visual pathways only to the lateral
geniculate. An intact flash response is expected with lesions of optic radiations and visual
cortex. Great care should be exercised in interpretation of latency abnormalities of the
flash VEP.
Ocular disorders
Many ocular disorders cause abnormalities in the PR-VEP. The effect of impaired visual
acuity and effective retinal illumination on the VEP have been discussed previously. The
VEP is not ordinarily used for the diagnosis of these disorders. Although some patients
with glaucoma have an increased latency and reduced amplitude of he P100, a normal
VEP cannot be interpreted as indicating normal intraocular pressure. The effects of ocular
and retinal disorders on the VEP are of interest only in the interpretation of VEPs used to
evaluate disorders a and behind the optic nerve.
Cortical blindness
Some patients with documented cortical blindness have normal PR-VEPs. The use of
smaller check size may help to bring out abnormalities, however, this is not routinely
done in most neurodiagnostic laboratories.
ECN 3/e - 188
Methods
Stimulus
Earphones
Headphones are placed on the ears for delivery of the auditory stimulus. For adults and
older children, large headphones which surround the eternal ear are used. These provide
for excellent sound delivery and block out much ambient noise. Young children and
infants cannot use these headphones because they may collapse the external auditory
canal, so small earphones are inserted into the EAC. Both types of phone should be
provided with the EP machine.
Types of stimulus
The EP apparatus delivers signals to the phone which produce one of a number of
different sounds. The most commonly used sounds are:
• Clicks
• Tones
• White noise
Clicks are routinely used as a stimulus for routine BAEP. The electrical signal is a
square wave with a rapid upstroke, plateau, and rapid down stroke to neutral voltage.
Since the direction of voltage change is opposite with the rising and falling phases of the
voltage signal, the earphone diaphragm moves toward the eardrum with one phase and
ECN 3/e - 189
away from the eardrum with the other. When the diaphragm moves toward the eardrum,
the air in the canal is compressed or condensed, therefore, this is termed condensation.
When the diaphragm moves away from the eardrum, the air in the canal is decompressed,
or rarefied. Therefore, this is termed rarefaction. The distinction is more than semantic,
the response can differ substantially between condensation and rarefaction clicks, so the
response is often averaged separately for the two modes. Of course, there has to be a
condensation movement of the diaphragm for every rarefaction, so both are delivered.
Interpretation of the BAEPs to rarefaction and condensation clicks is discussed below.
Table 15-1: BAEP Stimulus and recording parameters.
Parameter Value
Stimulus parameters
Rate 8-10/sec
Intensity 115-120 dB pe SPL
Duration of pulse 100 µsec
Stimulus character Monaural
Contralateral masking noise (60 dB pe SPL)
Stimulus polarity Rarefaction or condensation, summed
independently
Number of trials 1,000-4,000
Recording parameters
Recommended montages
Channel 1 Cz – Ai
Channel 2 Cz – Ac
LFF 10-30 Hz (-3 dB)
HFF 2,500-3,000 Hz (-3 dB)
Analysis time 10-15 msec
Measurements Wave I peak latency
Wave III peak latency
Wave V peak latency
Wave I amplitude
Wave V amplitude
Calculations I-III interpeak interval
III-V interpeak interval
I-V interpeak interval
Wave V/I amplitude ratio
Tones are produced by a sine wave, but are seldom used in routine BAEP. More
commonly, pure tones are used for audiometry, where hearing thresholds are determined
for different frequencies. Some disorders can produce a predominantly high-frequency
hearing loss, as with ototoxic drugs. This would be less well detected by conventional
BAEP than by pure tone audiometry.
White noise is not used as a stimulus, but is delivered to the non-stimulated ear as a
masking sound. This masking sound reduces spread of conduction of the stimulus sound
delivered to the opposite ear, so that the stimulation is not bilateral. Without this, air or
bone conduction of the stimulus can produce bilateral stimulation. The sound is similar to
that heard by a radio which is tuned to a frequency without a broadcasting station. Almost
every frequency is represented in the signal. The reason it is called white noise is because
ECN 3/e - 190
the broad spectrum of white noise is analogous to white light, where the white
appearance is due to the presence of all colors (frequencies or wavelengths) of light.
Stimulus rate
Clicks are delivered at a rate of 8-10/sec. This allows for reproducible identification of all
waves. Waves I, II, VI, and VII are reduced in amplitude at faster frequencies
Stimulus intensity
The Guidelines recommends stimulus intensities between 40 and 120 dB pe SPL. Many
EP machines do not give stimuli louder than that recommended for routine BAEP testing.
Intensity is set at 65 dB SL or HL. Reducing stimulus intensity is necessary only if
waveform identification is difficult. With decreasing stimulus intensity, waves II and VI
are reduced more than the other waves, allowing for more accurate identification of
waves I, III, and V.
Recording
Electrodes are placed in the following positions:
These electrode placements are identical to those for routine EEG recording.
• Channel 1 = Cz – Ai
• Channel 2 = Cz – Ac
where Ai is the ipsilateral ear and Ac is the contralateral ear. Therefore, for each ear
being stimulated, the first channel is from the ipsilateral ear in reference to the vertex and
the second channel is from the contralateral ear in reference to the vertex.
ECN 3/e - 191
Interpretation
Figure 15-1 shows a normal BAEP. The appearance can differ slightly, but the waves of
interest are all convex up and have the same multi-lobed appearance.
Waveform
identification
The waves routinely
analyzed in BAEP
testing are numbered I
through V. Waves VI
and VII are also
identified but not used
in interpretation.
Waves I and V should
be identified first.
Wave I is generated
Figure 15-1: Normal BAEP. by the distal portion of
Normal BAEP in response to stimulation of the right ear. Response from the acoustic nerve and
the right is on top, response from the left is on the bottom, note the is approximately 2 ms
absence of wave I in the contralateral recording. after the stimulus.
Wave I identification
is aided by recording from a contralateral electrode derivation, it is the only wave present
on ipsilateral but not contralateral recording.
Wave V may be generated by projections from the pons to the midbrain. There are
several criteria for identifying wave V. It normally appears at approximately 6 ms and is
often combined with wave IV into a single complex waveform. Wave V is also the first
waveform whose falling edge dips below the baseline.
The wave III-V complex has a wider separation with recording from the ipsilateral ear.
This means that the contralateral wave IV is slightly shorter latency and wave V is
slightly longer latency. Wave V is the last to disappear as stimulus intensity is decreased.
Wave III is thought to be generated by the projections from the superior olive through the
lateral lemniscus. It is the major peak between waves I and V.
Data analysis
ECN 3/e - 192
Latency is a more important measure than amplitude in the interpretation of BAEP data.
The most important measurements are:
• wave I latency
• wave III latency
• wave V latency
• Increased wave I latency: Damage to the most distal portion of the acoustic nerve.
Acoustic neuromas seldom affect wave I.
• Increased I – III interpeak interval: Defect in the pathway from the proximal eighth
nerve into the inferior pons. The lesion may be either in the nerve or in the brainstem.
This is the most common abnormality found in patients with acoustic neuromas.
• Increased III – V interpeak interval: Defect in the conduction between the caudal
pons and the midbrain.
• Increased I – III and wave III – V interpeak intervals: Lesion affects the brainstem at
and above the caudal pons with or without involvement of the acoustic nerve. In most
instances, this is due to a prominent lesion in the pons.
• Absence of wave I with normal III and V: May indicate a peripheral hearing disorder,
with the caveat that conduction in the caudal pons cannot be excluded.
• Absence of wave III with normal waves I and V: Normal, but if the wave I - V interval
is prolonged, then a lesion affecting conduction somewhere from the eighth nerve to
the midbrain is suspected.
• Absence of wave V with normal waves I and III: This is uncommon, but when present
indicates a lesion affecting the auditory pathways above the caudal pons. This is
considered an extreme prolongation of the wave III - V interval.
Abnormalities
Lesions of the lower brainstem or acoustic nerve can produce increased I-III interpeak
interval (Figure 15-2). This could be due to acoustic neuroma but is more likely to be due
to lesion at the cerebellopontine angle.
Lesion of the upper brainstem can produce increased III-V interpeak interval (Figure 15-
3). This can be due to stroke, mass lesions, or demyelinating disease. However, BAEP is
remarkable insensitive for diagnosis of MS.
ECN 3/e - 194
Pediatric
BAEPs
The most common use of
BAEPs in pediatric
patients is for assessment
of brainstem function in
premature infants.
Assessment of hearing is
also performed, and for
this, audiometry is
chiefly used.
Figure 15-3: Increased III-V interpeak interval of BAEP.
Patient with a lesion of the upper brainstem producing an increased III- Childhood
V interpeak interval and interfering with recording of wave V from the
contralateral side. BAEP in children is
mainly used to assess
hearing in patients who cannot cooperate with conventional hearing tests. An abnormal
BAEP is usually associated with abnormalities on behavioral testing of hearing, however,
a normal BAEP does not guarantee normal hearing. If the lesion is of the peripheral
auditory structures, threshold may be increased, but there may not be a change in the
wave I-V interpeak interval.
Neonates
Methods are slightly different in neonates than in older children. Earphones are used for
neonates rather than headphones, because the latter may collapse the external auditory
canal. Sedation may be required, and is best accomplished by chloral hydrate, although
meperidine plus secobarbital or meperidine plus promethazine are also used. Sedation
does not affect the short-latency EPs such as BAEP.
Wave I-V interpeak interval is increased in term newborns who have experienced
episodes of total asphyxia with subsequent damage to brainstem nuclei. The mortality
ECN 3/e - 195
and neurologic morbidity in such newborns is high. Newborns who have experienced
prolonged partial asphyxia sustain mainly hemispheric damage, and the wave –V interval
may be normal despite a poor neurologic outcome.
Clinical correlations
Acoustic neuroma
The most sensitive finding for the diagnosis of acoustic neuroma is prolongation of wave
I - III interpeak interval. If there is difficulty in obtaining wave I, the technician should
place an electrode in the external auditory canal for better recording. Alternatively,
electrocochleography can aid with identification of wave I. In patients with very large
tumors, there may be such severe damage that there are no reproducible waves after wave
I. In early cases, the BAEP has been abnormal, when imaging revealed an acoustic
neuroma. A sample finding with acoustic neuroma is shown in Figure 15-2.
Brainstem tumor
BAEP test results are abnormal in most patients with intrinsic tumors of the brainstem.
This is especially true in patients with pontine involvement. The usual abnormality is
delay or loss of waves III and V and increased wave I - III and wave III - V interpeak
interval.
Stroke
BAEP test results are abnormal in most patients with brainstem stroke. A few patients
with extensive brainstem infarctions have been reported to have normal BAEPs,
however. In some of these patients the amplitudes of the waveforms were low; however,
amplitude abnormalities are not emphasized in the interpretation of BAEPs.
Approximately 50% of patients with transient ischemic attacks affecting the posterior
circulation have latency abnormalities, and 50% of patients who recover from definite
brainstem strokes have normal BAEPs. A sample recording from a patient with a
brainstem lesion is shown in Figure 15-3.
Multiple sclerosis
BAEP testing is less sensitive than VEPs and SEPs for detection of clinically unsuspected
lesions in patients being evaluated for MS. The usual abnormalities are reduction in wave
V amplitude and increased wave III - V interpeak interval.
Most abnormalities are asymmetric, affecting the response from only one ear. Caution in
the interpretation of amplitude abnormalities is recommended. BAEP testing cannot
distinguish a demyelinating disease from tumors or infarction.
ECN 3/e - 196
BAEP test results are consistent with brain death if there are no reproducible waves after
wave I. Wave II may be intact in less than 10% of brain dead patients, reinforcing the
hypothesis that wave II is generated by the intracranial portion of the eighth nerve. The
presence of wave II is consistent with brain death in a patient who otherwise fulfills all
other clinical criteria and has no subsequent waves on the BAEP potential.
Audiometry
Audiometry uses BAEP to assess the function of the middle and inner ear rather than the
brainstem. The technique is similar to that used for conventional BAEP except that
special attention is given to the wave V latency as a function of stimulus intensity. With
increasing intensity, the wave V latency becomes progressively shorter. Stimuli are
delivered at stimulus intensities of 20, 40, 60, and 80 dB greater than threshold. Because
of the nature of the decibel scale, this is essentially a semilog plot.
The relationship between stimulus intensity and wave V latency (latency-intensity curve)
is linear in most individuals, with higher intensities producing shorter latencies.
Conductive hearing loss does not change the slope of this relationship but prolongs the
latency at each intensity. Therefore, the curve is shifted upward. The response looks as if
the intensities were turned down at every point, which is essentially what occurs with
conductive hearing loss. Sensorineural hearing loss produces a curve with two slopes. At
low intensities, there is decreased responsiveness of the end-organ, so that for a given
intensity the wave V latency is prolonged. With increases in intensity, there is more
recruitment of nerves than normal, so that the slope of the curve is steeper. At high
intensities, sufficient recruitment has occurred such that the latency may be normal. At
this point, the slope reverts to normal. This L-shaped curve is typical for sensorineural
hearing loss.
Audiometry is useful for evaluating patients for hearing loss when the localization of the
lesion is in doubt. Audiometry can also be used to follow patients receiving
chemotherapeutic agents which cause ototoxicity.
ECN 3/e - 198
• median
• ulnar
• peroneal
• tibial
Brief electric pulses are delivered to the peripheral nerve with the cathode proximal to the
anode. The stimulus cannot selectively activate sensory nerves, so a small muscle twitch
is seen. There are no effects of the retrograde motor volley in the motor nerves on central
projections of the sensory fibers.
Recording of the afferent nerve volley ensures that that the stimulus is adequate, and
determines whether there is a defect in peripheral conduction which would interfere with
interpretation of the central waveforms. Recordings from the spinal cord measure
electrical activity in white matter tracts and relay nuclei. Recordings from brain will
measure the projections from relay nuclei to the cortex. Short-latency responses are used
for clinical interpretation, long-latency responses are too variable to be helpful.
ECN 3/e - 199
The anatomy of the projections from the peripheral nerve to cortex is complex.
Unfortunately, there is not a precise anatomical correlate to each of the waves in the SEP.
This complicated interpretation, unfortunately.
SEPs are particularly useful for evaluation function of the spinal cord. Lesions of the cord
may be invisible to routine imaging, including MRI and myelography, yet may have
devastating effects on cord function. Transverse myelitis, multiple sclerosis, and cord
infarction are only three of the potential causes which can be missed on structural studies.
Median SEP
Methods
Stimulating electrodes are placed over the median nerve at the wrist with the cathode
proximal to the anode. Stimuli are square-wave pulses given at rates of 4-7/sec.
Erb’s point is 2-3 cm above the clavicle, just lateral to the attachment of the
sternocleidomastoid muscle. Stimulation at Erb’s point produces abduction of the arm
and flexion of the elbow. The second and fifth spinous processes are identified by
counting up from the seventh, notable by its prominence at the base of the neck. CPc and
CPi are scalp electrodes halfway between C3 and P3 or C4 and P4, where CPc is
contralateral to the stimulus and CPi is ipsilateral to the stimulus. These electrodes are
over the motor-sensory cortex. EPi is Erb’s point ipsilateral to the stimulus.
<< Table 16-2: Median nerve SEP stimulus and recording parameters. >>
Details of the recommended stimulus and recording parameters are presented in table x-x,
including analysis time, filter settings. These parameters are usually not changed for
individual studies. Number of trials averaged for adequate waveform identification is
500-2,000, but more may occasionally be needed. Duplicate trials overlayed on the
display helps greatly with waveform identification.
N9: The potential recorded from Erb’s point is sometimes called EP, but because of the
obvious abbreviation confusion, N9 is a preferable designation. The N9 potential is a
compound action potential from the axons stimulated y the median nerve stimulation.
While the N9 potential includes both orthograde sensory nerve potentials and retrograde
motor nerve potentials, all of the subsequent waves are due to sensory activation, alone.
ECN 3/e - 201
P14: The neck potentials include N13 and P14, with the latter used for clinical
interpretation. The origin is thought to be the caudal medial lemniscus.
N20: Scalp potentials include N18 and N20, but the latter is used for clinical
interpretation. Origin is thought to be the thalamocortical radiations.
Tibial SEP
Methods
Stimulus settings are similar to those used for median SEP, and details of the stimulus
and recording parameters are presented in Table 16-5. The proximal stimulating electrode
(cathode) is placed at the ankle between the medial malleolus and the Achilles tendon.
The anode is placed 3 cm distal to the cathode. A ground is placed proximal to the
stimulus electrodes, usually on the calf. Stimulus intensity is set so that each stimulus
produces a small amount of plantar flexion of the tows.
Table 16-4: SEP waveform origins.
Wave Origin
Median
N9/EP Afferent volley in plexus
N13 Dorsal horn neurons
P14 Caudal medial lemniscus
N18 Brainstem and thalamus?
N20 Thalamocortical radiations
Tibial
LP Dorsal roots and entry zone
N34 Brainstem and thalamus?
P37 Primary sensory cortex
• Channel 1: CPi-Fpz
• Channel 2: Cpz-Fpz
• Channel 3: Fpz-C5S
• Channel 4: T12S-Ref
LP: The lumbar potential (LP) is thought to arise from the afferent nerve volley in the
dorsal roots and dorsal root entry zone. Identification of the LP is usually easy. Patients
with peripheral neuropathy may have desynchronization of the afferent volley, such that
the amplitude of the nerve potentials may be low or inconsistent.
P37: P37 is a positive potential at about 37 ms which is seen from the scalp channels. The
origin is thought to be the primary sensory cortex. N34 precedes P37, but is not used for
clinical interpretation. Identification of the N34 and P37 is facilitated by overlapping
averaged traces, which is currently recommended to aid visual interpretation of EPs in
ECN 3/e - 204
general. N34 is the main negative wave in the FPz-C5S derivation and is preceded by a
small positive wave that is not used for interpretation (P31). The P37 is the major positive
wave in the Cpi-Fpz and CPz-Fpz channels. The amplitude may be very different
between these channels, reflecting an anatomic variance in waveform distribution.
Calculated data: The LP-P37 interval is the time from the cauda equina to the brain.
This is called the central conduction time (CCT).
Cervical spine data: Cervical potentials can occasionally be recorded during tibial SEP,
but these potentials are too inconsistent and too variable to be used for clinical
interpretation. Many laboratories do not even try to record these potentials, rather using
median SEP for clinical use.
Clinical correlations
Table 16-6 presents the expected SEP findings in various disorders.
Transverse myelitis
Transverse myelitis produces slowing of SEPs which depends on the site of the lesion.
Lesion in the lower cervical or thoracic cord increases central conduction time without
having an effect on brain conduction time. With recovery, the SEPs abnormalities are
improved, but may not return to normal.
Multiple sclerosis
SEP is abnormal in most patients with MS, and can be supporting evidence for a silent
lesion or confirmatory for a myelopathy. The most common finding in MS is an increase
in central conduction time of the tibial SEP with normal peripheral conduction (LP). This
is because the tibial SEP is assessing conduction along the longest myelinated nerve tract
of any of the evoked potentials. Brain conduction time of median nerve SEP is less
commonly increased than tibial nerve SEP CCT. A combined increase in BCT and CCT
can be due to tandem lesions, but also can be due to a single lesion in the cervical cord.
Peripheral neuropathy
Peripheral neuropathy slows peripheral conduction (N9 and LP) with normal BCT and
CCT. N9-P14 interval may be prolonged with lesions affecting the proximal portions of
the nerves, such as Guillain-Barre syndrome. GBS may also occasionally prolong CCT,
presumable by affecting the afferent nerve roots of the cauda equina.
B12 deficiency
Subacute combined degeneration from B12 deficiency delays or abolishes the cervical
and scalp SEPs. With treatment, the abnormalities improve, although not completely to
ECN 3/e - 206
normal. This parallels the clinical course, where there is improvement but also some
persistent deficit.
Brain death
Brain death is usually evaluated by EEG or blood flow studies, so the SEP is not typically
used as a confirmatory test. In brain death, the scalp potentials are absent, usually with
preservation of cervical potentials.
Stroke
SEPs are not commonly used for evaluation of stroke, but if performed, will show
attenuation, delay, and often absence of scalp potentials with stimulation of the limbs of
the affected side. Lesions of the motor-sensory cortical regions are much more likely to
produce abnormal SEPs than lesions elsewhere in the brain. In general, the severity of the
stroke deficit correlates with the degree of abnormality of the SEP, but wide variation is
common. SEP may be absent with subtle deficit and SEP may be preserved with major
deficit.
ECN 3/e - 208
Part V: Polysomnography
Chapter 17: Physiology of sleep and sleep disorders
Chapter 18: Sleep studies
ECN-CD: Table of Contents
ECN 3/e - 209
PSG is the recording of EEG and other physiological measures during sleep. The most
common indications for performance of the studies are:
Sleep stages were discussed in Chapter 4, Some of the information, below may be
duplicative, but is important to understanding of sleep physiology and pathology.
The exact mechanisms of sleep and wake onset are not known. Activity in the pontine
reticular formation, midbrain, and posterior hypothalamus are important for wakefulness.
Activity in the medullary reticular formation is important for generation of sleep. Sleep
and wake may be integrated in the basal forebrain.
Wakefulness is probably a function of tonic activity in cells that project to the cortex.
This activity increases neuronal excitability and may gate reactions to exogenous stimuli.
Sleep develops as an active process that is generated in sleep-promoting neurons, such as
the serotonergic raphe nuclei. The activation is probably promoted by a reduction in
exogenous and endogenous stimuli that indicates a need for sleep. The tonic activating
discharge and the response to exogenous stimuli are then suppressed, as are the patterned
spontaneous activities normally seen while awake.
Years of sleep deprivation experiments have not explained the need for sleep. One
proposed theory promotes the concept of sleep as a time for data management and
reorganization. During the waking state, the brain receives a great deal of information on
everything from music to tennis to physics. Much of this information is not ordered in a
conceptual format, and the brain cannot access the information in a structured way. For
example, there is a great difference between owning a tape of a lecture and understanding
ECN 3/e - 211
its content. Some data processing can occur in the waking state, but sleep may be
required to organize the day's input, integrate it with existing data, and perhaps discard
information that is seldom accessed or is judged by the brain to be useless or
uninteresting.
Sleep stages
Waking state
EEG in the waking state is discussed in Chapter 4 and summarized in Table 17-1. The
adult waking EEG consists of predominantly fast frequencies. When the eyes are closed,
a posterior dominant alpha rhythm predominates.
Sleep stage 1
Stage 1 is divided into stage 1A, light drowsiness, and stage 1B, deep drowsiness. Stage
1A shows desynchronization of the background with loss of the posterior alpha rhythm.
Theta activity is present but is not prominent.
Stage 1B is similar to stage 1A except that slow waves, mainly in the theta range, appear.
Vertex waves may be seen during this stage. Positive occipital sharp transients of sleep
are seen during this stage.
Sleep stage 2
Stage 2 is light sleep. For EEG purposes, we do not consider a study to include sleep
unless stage 2 is seen. The background consists of a mixture of frequencies. Delta activity
is present although not as prominent as in deeper stages of sleep. Theta and faster
frequencies are superimposed. Differentiation from stage 1B is made by the appearance
ECN 3/e - 212
of sleep spindles. Fusion of sleep spindles with vertex waves resul6ts in the K complex.
Vertex waves and K complexes are frequent.
Stage 3 sleep
Stages 3 and 4 are slow-wave sleep. Stage 3 is characterized by delta activity with a
frontal predominance. Sleep spindles, vertex waves, and K complexes persist but are not
as prominent as during stage 2. Mittens are seen during this stage and are composed of a
vertex wave fused to the end wave of a spindle. The small spindle wave is the thumb of
the mitten and the slow vertex wave is the hand.
Stage 4 sleep
ECN 3/e - 213
Stage 4 sleep is characterized by predominance of slow activity in the delta range. The
delta has a frontal predominance. Although some faster frequencies may be
superimposed, sleep spindles and vertex waves are seldom seen and, if present, are poorly
formed.
REM sleep
• Nocturnal PSG is indicated for patients who have clinical evidence of sleep apnea or
who have excessive daytime sleepiness. These conditions suggest a nocturnal sleep
disorder.
• MSLT is indicated when narcolepsy is suspected. MSLT testing cannot be used to
support the diagnosis of narcolepsy if the clinical history is not consistent. Excessive
daytime sleepiness, alone, is not an indication for MSLT.
ECN 3/e - 214
The Guidelines official recommendations list the following indications for sleep
monitoring:
Long-duration EEG monitoring of patients with suspected seizures usually does not
require all of the physiologic monitoring commonly performed during PSG studies; the
important information is EEG, ECG, and visual monitoring.
ECN 3/e - 215
Overview
Basic information is presented for understanding the performance and interpretation of
PSG including nocturnal PSG and MSLT. In-depth discussion of methods and
interpretation can be found in two excellent texts [insert these refs]
The technical recommendations for PSG are taken from the Guidelines in EEG, Evoked
Potentials, and Polysomnography (American Electroencephalographic Society, 1994).
These are referred to as the Guidelines.
Nocturnal polysomnography
Physiological measurements
The Guidelines recommends that the following physiologic measurements should be
made:
• EEG
• Electro-oculogram (EOG)
• Submental EMG
• ECG
• Respiration
• Blood oxygen saturation
• Expired CO2
• Body and limb movement
• Audiovisual monitoring
• Time
Not all laboratories record all of these parameters, however, the potential for error
increases with decreased information.
EEG
At least 6 channels of EEG should be recorded. The following electrode positions should
be used, as a minimum:
ECN 3/e - 216
Electrodes are attached using the same techniques described for routine EEG. Collodion
is more dependable for long-term recording than the use of electrode gel alone and is
therefore preferable for PSG.
Montages can be determined by the neurophysiologist, but in general should resemble the
montages use for routine EEG. Since most PSG equipment is computerized, nowadays,
individual channels are recorded and the montage can be selected during review.
Versions of the longitudinal bipolar and transverse bipolar montages are used, but only
one montage should be used to aid sleep-scoring. During interpretation, paper speed is
routinely set at 10 mm/sec, although faster speeds are occasionally helpful With digital
recording, this can be changed at the time of interpretation.
Electrooculogram (EOG)
Two channels are routinely used for EOG. Electrodes are placed in the following
positions:
The two channels are each eye in reference to the ipsilateral ear. Using this montage, eye
movements can be clearly differentiated from frontal slow activity. Eye movements will
produce potential of opposite polarity in the two eye leads. Frontal slow activity will
produce either slow waves of the same polarity or independent slow waves in the two
channels.
Submental EMG
ECN 3/e - 217
Submental EMG is recorded using standard cup electrodes placed underneath the chin.
The electrodes are connected to a standard EEG amplifier with the low-frequency filter
set at 10 Hz and the high-frequency filter set at 70 Hz. Gain is adjusted for each
individual but is in the same range as for EEG, about 7 µV/mm
Submental EMG is reduced in deep stages of sleep and virtually abolished in REM sleep.
The absence of EMG activity aids in identification of REM sleep.
Electrocardiogram (ECG)
ECG is recorded using two electrodes on the chest, usually the rostral sternum and left
lateral chest. Self-stick ECG electrodes are satisfactory. The low-frequency filter is set at
5 Hz and the high-frequency filter is set to 70 Hz. Gain is individualized but is usually
about 75 µV/mm
ECG recording serves two purposes. First, heart rate can change with respiratory distress,
so that patients with sleep apnea leading to hypoxia and hypercarbia may have initial
tachycardia followed by profound bradycardia. In this situation, ECG gives an estimate of
the severity of the apnea. The second purpose of ECG recording is to identify cardiac
artifact on EEG channels. This is less of a problem for PSG than for EEG performed for
epilepsy evaluation. Bipolar montages have less potential for ECG artifact than
referential montages.
Respiration
Respiratory monitoring is essential for diagnosis of sleep apnea. Measurements are made
of respiratory effort and airflow. Respiratory effort is recorded using thoracic and
abdominal strain gauge transducers, intercostal EMG, or thoracic and abdominal
impedance. Airflow is usually monitored using thermal sensory near the nares and mouth.
DC recordings are preferable, but AC recordings with a long time-constant are
acceptable.
Blood oxygenation
A pulse oximeter is most commonly used for measurement of oxygen saturation. Output
of the oximeter is fed to a reorder by a DC-amplifier, since absolute measurements
require steady-state DC recordings. The oximeter probe is usually on an earlobe but can
ECN 3/e - 218
be on a finger. Oximeters are fairly accurate, but may give falsely high readings in
patients with carbon monoxide in their blood, especially active smokers. Oximeters may
give falsely low results in patients with cool extremities or peripheral vascular disease.
Expired CO2
Expired CO2 can be measured by placing small sampling tubes below each nostril and
near the mouth. The chemical analyzer for CO2 is rapid and the output can be fed to the
recording device. The air at end expiration is largely alveolar, so that determination of
CO2 content is a fairly good indication of gas exchange. Patients with obstructive
disorders will have a fall-off in expired CO2 during the obstruction and may have higher
CO2 content after the obstruction.
Body movement
Surface EMG electrodes are placed over the tibialis anterior on one side for recording
EMG. This can reveal the presence of myoclonus and may aid in the diagnosis of restless
legs syndrome. Alternatively, accelerometers can be used. These are small devices which
produce a signal with a small amount of movement. However, because submental EMG
is already recorded, another EMG channel is much more convenient to use.
Audiovisual monitoring
Closed-circuit television (CCT) and microphones recording are used for monitoring of
behavior and movements. If the recording is not digital, split-screen recording of the EEG
and video signal is desirable, however, on digital systems, the streaming media is cued to
the EEG and physiologic display.
The camera is placed so that it can easily recording the patient in the bed. Modern
cameras are able to record in the very low-light conditions which are conducive to sleep.
Additional light may be provided from an infrared source. These light sources activate
detectors in the camera but will not wake the patient. Audio monitoring by small
microphones can detect vocalizations and be another indicator of respiratory effort.
The most important aspects of audiovisual monitoring are muscle twitches, signs of
arousal, axial and limb movements, respiratory effort, and seizures.
Time
Time is measured by a real-time clock in the computer, and is expressed as real and
elapsed time. Time of onset and cessation of recording should be noted on the record.
Paper records need to have accurate indicators of time, to associate with the audiovisual
ECN 3/e - 219
The following guidelines summarize the recommendations for a standard nocturnal PSG
recording:
• Make the room comfortable and quiet. Place the recording equipment in a separate
room.
• Begin he study as close to normal sleep time as possible.
• Minimize interruptions. Extra electrodes and sensors facilitate maintaining adequate
recording if the patient dislodges primary electrodes and sensors.
• Duration of recording should ideally be 8 hours, with 6.5 hours as a minimum.
Interpretation
Grading of nocturnal PSG is typically done in 20- to 40- second epochs. An epoch is
classified according to the predominant pattern during the epoch. For example, an epoch
characterized mainly by a desynchronized background may be classified as stage 1 sleep
even though there is some occasional posterior waking alpha activity.
Sleep onset is defined as either the first of three contiguous epochs of stage 1 sleep or the
first of any stage 2, 3, or 4 sleep. Three consecutive epochs of state 2, 3, or 4 sleep are not
required. With simultaneous monitoring of many physiologic variables, the amount of
generated data can be overwhelming. The Guidelines recommends the following sleep
measurements.
Sleep efficiency is the percent of total time in bed spent asleep. In addition, graphs of
sleep stage progression are drawn. These are usually generated automatically by the
acquisition software.
• Mean and range of heart rate during wake and sleep states
• Arrhythmias, if present
• Cardiac response to respiratory changes, e.g. apnea
EMG data are analyzed for myoclonus and differentiation is made between myoclonus
associated with arousal, myoclonus associated with epileptiform activity on EEG, and
myoclonus not associated with other physiologic changes.
Disorders
Sleep disorders can be classified according to the following categories:
ECN 3/e - 221
• Hypersomnias
• Insomnias
• Disorders of the sleep-wake cycle
• Arousal and paroxysmal disorders in sleep
• Excessive daytime sleepiness
Narcolepsy
Narcolepsy is characterized by daytime sleep attacks. There are two basic types of
narcolepsy:
Non-REM narcolepsy has non-REM sleep during attacks. Night sleep is normal. REM-
narcolepsy has REM sleep during sleep attacks. At night, there is sleep fragmentation and
shortened REM latency.
The MSLT may detect short REM latency or sleep-onset REM. The MSLT tests naps but
may catch a narcoleptic sleep attack.
Sleep apnea
Sleep apnea is probably the most common clinical reason for ordering PSG. There are
three basic types of sleep apnea:
All types are characterized by loss of airflow for 10 seconds or more. Patients with OSA
continue to have respiratory effort that gradually increases because the movements are
ineffectual. Eventually, partial arousal results in opening of upper airway passages and
restoration of ventilation. Patients with central sleep apnea lose air movement because of
loss of respiratory drive. With subsequent hypoxia and hypercarbia, there is partial
arousal and restoration of normal ventilation.
Methods
Conventional EEG electrodes are used, and a waking recording is made. The patient is
then asked to go to sleep. The technician marks the time on the record.
1) Patient has a normal night’s sleep prior to the recording. Most neurophysiologists
believe that it is important to have the patient under PSG study to evaluate the quality
of the previous night’s sleep. This helps to determine whether a positive MSLT might
be caused by a disorder of nocturnal sleep, rather than being primary. A PSG
recording is probably not necessary in all patients.
2) Electrodes are placed according to the 10-20 Electrode Placement system. The entire
array is probably not necessary, however, it is easily placed in most EEG laboratories.
If a limited array is placed, central and occipital leads are essential for identification
of central vertex activity and the posterior dominant rhythm. In addition to EEG
leads, electrodes should be placed for monitoring the following physiologic
parameters.
a) EOG
b) Submental EMG
c) ECG
3) At least four naps are begun at scheduled intervals. The technician lowers the lights
and asks the patient to go to sleep. Approximately 15 minutes are recorded before the
first nap. After the “goodnight” command, recordings are made until the following
criteria are met.
a) 20 minutes without sleep
b) 15 minutes of continuous sleep
c) 20 minutes of interrupted sleep, even if less than 15 minutes of sleep occurred.
Interpretation
ECN 3/e - 223
The most characteristic abnormality found with sleep latency tests is short sleep latency.
A mean sleep latency of less than 5 minutes is virtually diagnostic of hypersomnolence.
Latency of 10 minutes or more is normal. A mean sleep latency between 5 and 10
minutes is borderline. The report may indicate that consistent mean sleep latency of less
than 10 minutes is suggestive of a sleep disorder but is not diagnostic.
• Sleep deprivation
• Certain medications, especially sedatives, antihistamines, and stimulants
• Withdrawal of some medications
• Age
Many patients with excessive daytime sleepiness (EDS) will have shorter sleep latencies.
Patients with narcolepsy will often have sleep-onset REM periods. The interpreter should
be sure that the patient is not sleep-deprived before making the conclusion of sleep-onset
or short-latency REM periods.
ECN 3/e - 224
Abbreviations
µsec: microsecond
ACh: acetylcholine
AChE: acetylcholinesterase
AChR: acetylcholine receptor
AIDP: acute inflammatory demyelinating polyradiculoneuropathy
ALS: amyotrophic lateral sclerosis
BAEP: brainstem auditory evoked potential
BD: brain death
CIDP: chronic inflammatory demyelinating polyneuropathy
CMAP: compound motor action potential
DL: distal latency
EEG: electroencephalography
EMF: electromotive force
EMG: electromyography; usually implies the needle study although surface and single-
fiber studies are also included in this term
EP: evoked potential
HFF: high-frequency filter
HMSN: hereditary motor sensory neuropathy
Hz: hertz
LB: longitudinal bipolar (montage)
LEMS: Lambert-Eaton myasthenic syndrome
LFF: low-frequency filter
m: meter
MEP: motor evoked potential
MG: myasthenia gravis
MMN: multifocal motor neuropathy
MS: multiple sclerosis
msec: millisecond
NCS nerve conduction study. Includes assessment of nerve conduction velocity,
amplitude, and waveform.
NCV: nerve conduction velocity
Ref: reference (electrode or montage)
sec: second
SEP: somatosensory evoked potential
SNAP: sensory nerve action potential
TB: transverse bipolar (montage)
TM: transverse myelitis
VEP: visual evoked potential