F Wave

F waves were named in reference to “foot” since they were originally recorded in small foot muscles, though they may be generated by the stimulation of any motor or mixed nerve. A supramaximal stimulus applied at any point along the course of a motor nerve elicits an F wave in a distal muscle that follows the CMAP (M response). The impulse travels antidromically to the spinal cord and the F wave is produced by backfiring of motor neurons. An average of 5–10% of the motor neurons available in the motor neuron pool backfire after each stimulus. The afferent and efferent loops of the F wave are motor with no intervening synapse. Hence, the F waves test the integrity of the entire motor axons, including the ventral roots.

The F waves are low-amplitude and ubiquitous responses that are typically variable in latency, amplitude, and morphology (Figure 3-1A and B). Their variability is explained by differing groups of motor neurons generating the recurring discharge with each individual group of neurons having different number of motor neurons and conducting properties. Several parameters may be analyzed, but the minimal F wave latency is the most reliable and useful measurement since it represents conduction of the largest and fastest motor fibers. Since F wave latencies vary from one stimulus to the next, an adequate study requires that about 10 F waves be clearly identified. The minimal F wave latency is also dependent on the length of motor axons which correlates with the patient’s height and limb length. The most sensitive criterion of abnormality in a unilateral disorder is a latency difference between the two sides, or between two similar nerves in the same limb. Absolute latencies are most useful only for sequential reassessment of the same nerve. F wave persistence is a measure of the number of F waves obtained for the number of stimulations. This varies between individuals and is inhibited by muscle activity while it is enhanced by relaxation or the use of Jendrasik maneuver. It is usually above 50% except when stimulating the peroneal nerve while recording the extensor digitorum brevis. F wave chronodispersion is the degree of scatter among consecutive F waves and is determined by the difference between the minimal and maximal F wave latencies. It indicates the range of motor conduction velocities between the smallest and largest myelinated motor axon in the nerve. The F wave conduction velocity may be calculated after distal and proximal supramaximal stimulations and provides a better comparison between proximal and distal (forearm or leg) segments.

F wave latencies are prolonged in most polyneuropathies, particularly the demyelinating type, including Guillain-Barré syndrome and chronic inflammatory demyelinating polyneuropathy (CIDP) (Figure 3-1C and D). F wave latencies in radiculopathies have a limited use. They may be normal despite partial motor axonal loss because the surviving axons conduct normally, and in single radiculopathies since most muscles have multiple root innervation. Finally, focal slowing at the root level may get diluted by the relatively long motor axon.

A Wave

The A wave (axonal wave) is a potential that is seen occasionally during recording of F waves at supramaximal stimulation. The A wave follows the CMAP and often precedes, but occasionally follows, the F wave. The A wave may be mistakenly considered for an F response but its constant latency and morphology in at least 10 out of 20 stimulations differentiates it from the highly variable morphology and latency of the F wave (Table 3-1).

The A wave may be seen in up to 5% of asymptomatic individuals, particularly while studying the tibial nerve (Figure 3-2A and B). In contrast, recording multiple or complex A waves from several nerves is often associated with acquired or inherited demyelinating polyneuropathies (Figure 3-3). A waves are sometimes seen in axon-loss polyneuropathies, motor neuron disease, and radiculopathies. The exact pathway of the A wave is unknown but it may be generated as a result of ephaptic transmission between two axons with the action potential conducting back down the nerve fiber to the muscle. The A wave may also appear following sprouting and reinnervation along the examined nerve. The constant morphology and latency of the A wave is best explained by the fixed point of a collateral sprout or ephapse. When the A wave follows rather than precedes the F response, it suggests that the regenerating collateral fibers are conducting very slowly.

Figure 3-2 Tibial A wave, stimulating the tibial nerve at the ankle while recording the abductor hallucis muscle. (A) Waveforms shown in a raster mode. (B) Waveforms superimposed. Note that the response has a constant morphology and latency, which results in perfect superimposition.

Figure 3-3 Complex A wave stimulating the median nerve at the wrist while recording abductor pollicis brevis in a patient with Guillian-Barré syndrome shown in a raster form. Note that the complex response has a constant morphology and latency.

H Reflex

The H reflex, named after Hoffmann for his original description, is an electrical counterpart of the stretch reflex which is elicited by a mechanical tap. Group 1A sensory fibers constitute the afferent arc which monosynaptically or oligosynaptically activate the alpha motor neurons that in turn generate the efferent arc of the reflex through their motor axons. The H reflex amplitude may be occasionally as high as the M amplitude but it is often lower with the H/M amplitude ratio usually not exceeding 0.75.

The H reflex and F wave can be distinguished by increasing stimulus intensity (see Table 3-1). The H reflex is best elicited by a long-duration stimulus which is submaximal to produce an M response, whereas the F wave requires supramaximal stimulus Also, the F wave can be elicited from any limb muscle while the H reflex is most reproducible with stimulating the tibial nerve while recording the soleus muscle which assess the integrity of the S1 arc reflex and is equivalent to the Achilles reflex (Figure 3-4). Finally, the H reflex latency (and often amplitude) is constant when elicited by the same stimulus intensity, since it reflects activation of the same motor neuron pool.

Figure 3-4 H reflex stimulating the tibial nerve at the popliteal fossa while recording the soleus/gastrocnemius muscles. (A) Waveforms in a raster form. (B) Waveforms superimposed. Note that the response appears with submaximal stimulations and disappears with supramaximal stimulations.

The H reflex is most useful as an adjunct study in the diagnosis of peripheral polyneuropathy or S1 radiculopathy. The H reflex latency and amplitude is the most sensitive, yet nonspecific, among the nerve conduction studies in the early phases of Guillain-Barré syndrome. The H reflex may be absent in healthy elderly subjects and isolated abnormalities of the H reflex are nondiagnostic since they may reflect pathology anywhere along the reflex arc.

Blink Reflex

The blink reflex generally assesses the facial and trigeminal nerves and their connections within the pons and medulla. It has an afferent limb, mediated by sensory fibers of the supraorbital branch of the ophthalmic division of the trigeminal nerve, and an efferent limb mediated by motor fibers of the facial nerve and its superior motor branches.

The supraorbital nerve is stimulated over the suprarorbital notch and the blink responses are recorded from the orbicularis oculi bilaterally using a two channel recording setting. The blink reflex has two components, an early R1 and a late R2 response. The R1 response is present only ipsilateral to the stimulation and is usually a simple triphasic waveform with a di-synaptic pathway between the main trigeminal sensory nucleus in the midpons and the ipsilateral facial nucleus in the lower pontine tegmentum. The R2 response is a complex waveform and is the electrical counterpart of the corneal reflex. It is typically present bilaterally with an oligosynaptic pathway between the nucleus of the trigeminal spinal tract in the ipsilateral pons and medulla, and interneurons forming connections to the ipsilateral and contralateral facial nuclei.

The blink reflex is most useful in unilateral lesions such as facial palsy, trigeminal neuropathy, or lower brainstem lesion. With facial nerve lesions, the R1 and R2 potentials are absent or delayed with supraorbital stimulation ipsilateral to the lesion, while the R2 response on the contralateral side is normal. With trigeminal nerve lesions, the ipsilateral R1 and R2, contralateral R2 are absent or delayed with ipsilateral stimulation while all responses are normal with contralateral stimulation. With a midpontine lesion involving the main sensory trigeminal nucleus or the pontine interneurons to the ipsilateral facial nerve nucleus or both, supraorbital stimulation on the side of the lesion results in an absent or delayed R1, but an intact ipsilateral and contralateral R2. Finally, with a medullary lesion involving the spinal tract and trigeminal nucleus or the medullary interneurons to the ipsilateral facial nerve nucleus or both, supraorbital stimulation on the affected side results in a normal R1 and contralateral R2, but an absent or delayed ipsilateral R2. In demyelinating polyneuropathies such as the Guillain-Barré syndrome or Charcot-Marie-Tooth disease type 1, the R1 and R2 responses may be markedly delayed, reflecting slowing of motor fibers, sensory fibers, or both.

Physiology and Principles

A quantum is the amount on acetylcholine packaged in a single vesicle, which contains approximately 5000 to 10 000 acetylcholine molecules. Each quantum (vesicle) released results in a 1 mV change of postsynaptic membrane potential. During rest, spontaneous release of individual quanta forms the basis of miniature endplate potential.

The number of quanta released after a nerve action potential depends on the number of quanta in the immediately available (primary) store and the probability of release, i.e., m = p × n, where m = the number of quanta released during each stimulation, p = the probability of release (effectively proportional to the concentration of calcium and typically about 0.2, or 20%), and n = the number of quanta in the immediately available store. In normal conditions, a single nerve action potential triggers the release of 50–300 vesicles (quanta) with an average equivalent to about 60 quanta (60 vesicles). In addition to the immediately available store of acetylcholine located beneath the presynaptic nerve terminal membrane, a secondary (or mobilization) store starts to replenish the immediately available store after 1–2 seconds of repetitive nerve action potentials. A large tertiary (or reserve) store is also available in the axon and cell body.

The end plate potential is the potential generated at the postsynaptic membrane following a nerve action potential. Since each vesicle released causes a 1 mV change in the postsynaptic membrane potential, this results in about 60 mV change in the amplitude of the membrane potential.

In normal conditions, the number of quanta (vesicles) released at the neuromuscular junction by the presynaptic terminal far exceeds the postsynaptic membrane potential change required to reach the threshold needed to generate a postsynaptic muscle action potential. This safety factor ensures an endplate potential that is always above threshold and results in muscle fiber action potential. In addition to quantal release, several other factors contribute to the safety factor and endplate potential including acetylcholine receptor conduction properties, acetylcholine receptor density, and acetylcholinesterase activity.

Following depolarization of the presynaptic terminal, voltage-gated calcium channels open leading to calcium influx. Through a calcium-dependent intracellular cascade, vesicles are docked into active release zones and acetylcholine molecules are released. Calcium then diffuses slowly out of the presynaptic terminal in 100–200 ms.

The rate at which motor nerves are repetitively stimulated dictates whether calcium accumulation plays a role in enhancing the release of acetylcholine. At slow rate of RNS (i.e., a stimulus than every 200 ms or more, or a stimulation rate of <5 Hz), the role of calcium in acetylcholine release is not enhanced and subsequent nerve action potentials reach the nerve terminal long after calcium has dispersed. In contrast, with rapid RNS (i.e., a stimulus every 100 ms or less, or stimulation rate >10 Hz), calcium influx is greatly enhanced and the probability of release of acetylcholine quanta increases.

The surface-recorded CMAP obtained during routine motor NCS represents the summation of all muscle fiber action potential generated in a muscle following supramaximal stimulation of all motor axons.

Slow Repetitive Nerve Stimulation

Slow RNS is usually performed by applying 3–5 supramaximal stimuli to a mixed or motor nerve at a rate of 2–3 Hz. This rate is low enough to prevent calcium accumulation, but high enough to deplete the quanta in the immediately available store before the mobilization store starts to replenish it. A total of 3–5 stimuli are adequate since the maximal decrease in acetylcholine release occurs during the first 3-5 stimuli.

Calculation of the decrement with slow RNS requires comparing the baseline CMAP amplitude to the lowest CMAP amplitude (usually the third or fourth). The CMAP decrement is expressed as a percentage and calculated as follows:

In normal conditions, slow RNS does not cause a CMAP decrement. Although the second to fifth endplate potentials fall in amplitude, they remain above threshold (due to the normal safety factor) and ensure generation of muscle fiber action potential with each stimulation. In addition, the secondary store begins to replace the depleted quanta after the first few seconds with a subsequent rise in the endplate potential. Hence, all muscle fibers generate muscle fiber action potentials and the CMAP does not change (Figures 3-5 and 3-6A). In postsynaptic neuromuscular junction disorders (such as myasthenia gravis), the safety factor is reduced since there are fewer available acetylcholine receptors. Hence, the baseline endplate potential is reduced but usually still above threshold. Slow RNS results in a decrease in endplate potential amplitudes at many neuromuscular junctions. As endplate potentials become subthreshold, there is a decline in the number of muscle fiber action potentials, leading to a CMAP decrement (see Figures 3-5 and 3-6B). In presynaptic disorders (such as Lambert-Eaton myasthenic syndrome), the baseline end plate potential is low, with many endplates not reaching threshold. Hence, many muscle fibers do not fire, resulting in a low baseline CMAP amplitude. With slow RNS, there is further CMAP decrement, due to the further decline of acetylcholine release with the subsequent stimuli, resulting in further loss of many endplate potentials and muscle fiber action potentials (see Figures 3-5 and 3-6A).

Figure 3-5 Slow repetitive stimulation effect on end plate potential (EPP), single fiber action potential (SFAP), also referred as muscle action potential (MAP), and compound muscle action potential (CMAP) in normal, myasthenia gravis (MG) and Lambert-Eaton syndrome (ELS).

(Adapted from Oh S. Clinical electromyography, neuromuscular transmission studies. Baltimore, MD: Williams and Wilkins, 1988, with permission.)

Figure 3-6 Slow repetitive nerve stimulation at 3 Hz of the median nerve at the wrist while recoding the abductor pollicis brevis. (A) Normal. (B) Patient with severe generalized myasthenia gravis showing a significant CMAP decrement. (C) Patient with Lambert-Eaton myasthenic syndrome showing also a significant CMAP decrement. Note that the baseline (first) CMAP in myasthenia gravis is normal (12 mV) while it is low in amplitude in the Lambert-Eaton myasthenic syndrome (3.5 mV).

Rapid Repetitive Nerve Stimulation

Rapid RNS is most useful in patients with suspected presynaptic neuromuscular junction disorders such as Lambert-Eaton myasthenic syndrome or botulism. The optimal frequency is 20–50 Hz for 2–10 seconds. A typical rapid RNS applies 200 stimuli at a rate of 50 Hz (i.e., 50 Hz for 4 seconds). Calculation of CMAP increment after rapid RNS is as follows:

With rapid RNS or postexercise CMAP evaluation, there are two competing forces that are acting on the nerve terminal. First, stimulation tends to deplete the pool of readily available synaptic vesicles. This depletion reduces transmitter release by reduction of the number of vesicles that are released in response to a nerve terminal action potential. Second, calcium accumulates within the nerve terminal, thereby increasing the probability of synaptic vesicle release. In a normal nerve terminal, the effect of depletion of readily available synaptic vesicles usually predominates, so that with rapid RNS, the number of vesicles released decreases. However, the endplate potential does not fall below threshold due to the safety factor. Hence, the supramaximal stimulus generate muscle fiber action potentials at all endplates and no CMAP decrement occurs (Figures 3-7 and 3-8A).

Figure 3-7 Rapid repetitive stimulation effect on end plate potential (EPP), single fiber action potential (SFAP), also referred as muscle action potential (MAP), and compound muscle action potential (CMAP) in normal, myasthenia gravis (MG) and Lambert-Eaton syndrome (ELS).

(Adapted from Oh S. Clinical electromyography, neuromuscular transmission studies. Baltimore, MD: Williams and Wilkins, 1988, with permission.)

Figure 3-8 Rapid repetitive stimulation at 50 Hz of the median nerve at the wrist while recoding the abductor pollicis brevis. (A) Normal. (B) Patient with Lambert-Eaton myasthenic syndrome showing significant increment of the CMAP.

A brief (10 second) period of maximal voluntary isometric exercise is much less painful and has the same effect as rapid RNS at 20–50 Hz. A single supramaximal stimulus is applied to generate a baseline CMAP. Then, the patient performs a 10 second maximal isometric voluntary contraction which is followed by another stimulus and a postexercise CMAP.

In a presynaptic disorder (such as Lambert-Eaton myasthenic syndrome), very few vesicles are released and many muscle fibers do not reach threshold, resulting in low baseline CMAP amplitude. With rapid RNS, the calcium concentrations in the nerve terminal can rise high enough to stimulate synaptic vesicle fusion for a sufficient number of synaptic vesicles to result in an endplate potential capable of action potential generation. This leads to many muscle fibers reaching threshold and firing and results in a CMAP increment (see Figures 3-7, 3-8B and 3-9). The increment is typically higher than 200% in Lambert-Eaton myasthenic syndrome, and is usually 30–100% in patients with botulism (Table 3-2).

Figure 3-9 Pre- and postexercise CMAPs of the ulnar nerve following stimulation at the wrist while recoding the abductor digiti minimi in a patient with Lambert-Eaton myasthenic syndrome. The responses are superimposed. Note that the CMAP at rest (baseline) is low in amplitude but there is a prominent (about 200%) increment after 10 seconds of exercise.

In a postsynaptic disorder (such as myasthenia gravis), rapid RNS causes no change of CMAP since the depleted stores are compensated by the calcium influx. In severe postsynaptic blockade (such as during myasthenic crisis), the increased quantal release cannot compensate for the marked neuromuscular block resulting in a drop in endplate potential amplitude. Hence, fewer muscle fiber action potentials are generated with an associated CMAP decrement (see Table 3-2).

Fiber Density

Fiber density is determined by the number of single fiber potentials firing almost synchronously with the initially identified single fiber potential. Increased muscle fiber clustering indicates collateral sprouting. Simultaneously firing single fiber potentials within 5 ms after the triggering single fiber unit are counted at 20–30 sites. For example, in the normal extensor digitorum communis muscle, single fibers fire without nearby discharges in 65–70% of random insertions; with only two fibers discharging in 30–35%; and with three fibers discharging in 5% or fewer. An average number of single muscle fiber potentials per recording site can be calculated. In conditions with loss of mosaic distribution of muscle fibers from a motor unit, such as reinnervation, fiber density increases.

Jitter

Jitter analysis is most useful in the assessment of neuromuscular junction disorders. The jitter represents the variability of the time interval between two muscle fiber action potentials (muscle pair) that are innervated by the same motor unit. In other words, it is the variability of the interpotential intervals between repetitively firing paired single fiber potentials (Figure 3-10). Jitter can be determined by using a commercially available computer software program. It is calculated as the mean value of consecutive interval difference, over a number of 50–100 discharges as follows:

Figure 3-10 Jitter single fiber EMG study of the frontalis muscle shown in a superimposed mode. The first potential is the triggered potential while the second is the slave potential. (A) Normal. (B) Patient with myasthenia gravis. Note the significant jitter of the slave potential (B) compared to (A).

where MCD is the mean consecutive difference, IPI 1 is the interpotential interval of the first discharge, IPI 2 of the second discharge, etc., and N is the number of discharges recorded.

Neuromuscular blocking is the intermittent failure of transmission of one of the two muscle fiber potentials. This reflects the failure of one of the muscle fibers to transmit an action potential due to the failure of the endplate potential to reach threshold. Blocking represent the most extreme abnormality of the jitter and is measured as the percentage of discharges of a motor unit in which a single fiber potential does not fire. For example, in 100 discharges of the pair, if a single potential is missing 30 times, the blocking is 30%. In general, blocking occurs when the jitter values are significantly abnormal.

The results of single fiber EMG jitter study are expressed by: (1) the mean jitter of all potential pairs, (2) the percentage of pairs with blocking, and (3) the percentage of pairs with normal jitter. Since jitter may be abnormal in 1 of 20 recorded potentials in healthy subjects, the study is considered to indicate defective neuromuscular transmission if (1) the mean jitter value exceeds the upper limit of the normal jitter value for that muscle, (2) more than 10% (more than two pairs) exhibits jitter values above the upper limit of the normal jitter, or (3) there is any neuromuscular blocking.

Jitter analysis is highly sensitive but not specific. Although it is frequently abnormal in myasthenia gravis and other neuromuscular junction disorders (see Figure 3-10), it may also be abnormal in a variety of neuromuscular disorders including motor neuron disease, neuropathies, and myopathies. Thus, the diagnostic value of jitter has to be considered in the contest of the patient’s clinical manifestations, and routine electrodiagnostic findings.