Clinical electromyography is a distinct medical discipline that plays a pivotal role in the diagnosis of neuromuscular disorders (Katirji et al., 2002). The designations clinical electromyography, electrodiagnostic examination, and electroneuromyography are used interchangeably to encompass the electrophysiological study of nerve and muscle; the terms needle electromyography or needle electrode examination are reserved for the specific testing that involves needle electrode evaluation of muscle. Although many still refer to all such testing as simply electromyography (EMG), use of the word without a descriptor is discouraged because it can be confusing, often implying only the needle electrode part of the evaluation. For clarity, this chapter will refer to clinical EMG or needle EMG in the context of the discussion.

The clinical EMG examination is an important diagnostic tool that helps localize a neuromuscular problem at the motor neurons, nerve roots, peripheral nerves, neuromuscular junction, or muscle. It also helps establish the underlying process in these disorders and assess their management and prognosis. Electrodiagnostic testing provides the most valuable diagnostic information when the clinical assessment suggests a short list of differential diagnoses. The clinician should perform a detailed or focused neurological examination before referring the patient for a clinical EMG, which in turn serves as an independent procedure to provide an objective assessment of the peripheral nervous system (PNS) (Katirji, 2002). Patients with complex clinical pictures are best served by a neurological consultation prior to ordering electrodiagnostic testing.

The clinical EMG examination is composed of two main tests: nerve conduction studies (NCS) and needle EMG. These tests complement each other, and both often are necessary for a definite diagnosis. Additional electrodiagnostic procedures include assessment of F waves, H reflexes, and blink reflexes; repetitive nerve stimulation; and single-fiber EMG. A focused history and examination will help the electromyographer design the most appropriate electrodiagnostic study (Preston and Shapiro, 2005). The electromyographer must be proficient in using modern electrodiagnostic equipment and applying electrodiagnostic techniques, know the normal values for commonly and uncommonly examined nerve conduction studies and for motor unit action potentials (MUAPs) in different muscles, and be familiar with the specific and nonspecific electrodiagnostic findings in different neuromuscular disorders.

Nerve Conduction Studies

Principles

Electrical stimulation of nerve fibers initiates impulses that travel along motor, sensory, or mixed nerves and evoke a compound action potential. The three types of NCS are motor, sensory, and mixed. Analysis of the compound muscle action potential (CMAP), evoked by stimulating a nerve while recording from a muscle, indirectly assesses the conduction characteristics of motor fibers. Analysis of the sensory nerve action potential (SNAP) assesses the sensory fibers by stimulating a nerve and recording directly from a cutaneous nerve. Mixed NCS directly assess the sensory and motor fibers simultaneously by stimulating and recording from a mixed nerve and analyzing the mixed nerve action potential (MNAP). Use of standard NCS allows precise lesion localization and accurate characterization of peripheral nerve function.

Motor Nerve Conduction Studies

The performance of motor NCS requires stimulating a motor or mixed peripheral nerve while recording the CMAP from a muscle innervated by that nerve. A pair of recording electrodes consists of an active lead, G1, placed on the belly of the muscle and a reference (indifferent or inactive) lead, G2, on the tendon (belly-tendon recording). The propagating muscle action potential, originating under G1 located near the motor point, gives rise to a simple biphasic waveform with an initial negativity. Initial positivity suggests incorrect positioning of the active electrode away from the motor end-plate zone or a volume-conducted potential from distant muscles activated by anomalous innervation or by accidental spread of stimulation to other neighboring nerves.

The nerve usually is stimulated, whenever technically feasible, at two or more points along its course. Shorter nerves such as the axillary, femoral, and facial nerves are stimulated at only one point, because the more proximal portions of the nerves are inaccessible. Otherwise, the nerve typically is stimulated distally near the recording electrode and more proximally to evaluate one or more proximal segments. Motor NCS evaluate several measurements (Fig. 32B.1):

CMAP amplitude: The usual measure of amplitude is from baseline to negative peak and expressed in millivolts. When recorded with surface electrodes, CMAP amplitude is a semiquantitative measure of the number of axons conducting between the stimulating and recording points. CMAP amplitude also depends on the relative conduction speed of the axons, the integrity of the neuromuscular junctions, and the number of muscle fibers that are able to generate action potentials.

CMAP duration: This measurement usually is the duration of the negative phase of the evoked potential and expressed in milliseconds. It is a function of the conduction rates of the various axons forming the examined nerve and the distance between the stimulation and recording electrodes. The CMAP generated from proximal stimulation is slightly longer in duration and of lower amplitude than that obtained from distal stimulation, as a result of temporal dispersion and phase cancellation (see forthcoming section).

CMAP area: This usually is limited to the negative phase area under the waveform and shows linear correlation with the product of amplitude and duration. Measurement is in millivolts per millisecond and requires electronic integration using computerized equipment. The ability to measure CMAP area has practically replaced the need to measure its duration.

Latencies: Latency is the time interval between nerve stimulation (shock artifact) and the CMAP onset. Expression of latency is in milliseconds and reflects the conduction rate of the fastest-conducting axon. Whenever technically possible, the nerve is stimulated at two points: a distal point near the recording site and a more proximal point; the measures obtained are the distal latency and proximal latency, respectively. Both latencies depend mostly on the length of the nerve segment and, to a much lesser extent, on neuromuscular transmission time and propagation time along the muscle membrane.

Conduction velocity: This is a computed measurement of the speed of conduction expressed in meters per second. Measurement of conduction velocity allows comparison of the speed of conduction of the fastest fibers between different nerves and subjects, irrespective of the length of the nerve. The calculation requires measurement of the length of the nerve segment between distal and proximal stimulation sites. Measuring the surface distance along the course of the nerve estimates the nerve length; it should be more than 10 cm to improve the accuracy of surface measurement.

Fig. 32B.1 Motor nerve conduction study of the median nerve, revealing a typical compound muscle action potential (CMAP) with distal (wrist) and proximal (elbow) stimulations, showing the distal and proximal latencies and CMAP amplitudes, durations, and areas. The proximal CMAP has a lower amplitude (12.6 mV versus 11.3 mV) and area (37.3 mV/ms versus 34.50 mV/ms) than the distal CMAP because of temporal dispersion and phase cancellation. The proximal conduction velocity is calculated by measuring the distance of the elbow-to-wrist segment and using the following formula:

So for the conduction velocity in this example, 210 mm/6.9 msec − 3.5 msec = 62 m/second.

As with latencies, motor conduction velocity measures the speed of conduction of the fastest axon. In contrast with motor latency, however, motor nerve conduction velocity is a pure nerve conduction time, because neuromuscular transmission time and muscle fiber propagation time are common to both stimulation sites, and the latency difference between two points is the time required for the nerve impulse to travel from one stimulus point to the other.

Sensory Nerve Conduction Studies

Sensory axons are evaluated by stimulating a nerve while recording the transmitted potential from the same nerve at a different site. Therefore, SNAPs are true nerve action potentials. The measurement of antidromic sensory NCS requires recording potentials directed toward the sensory receptors. Obtaining orthodromic responses requires recording potentials directed away from these receptors. Sensory latencies and conduction velocities are identical with either method, but SNAP amplitudes generally are higher in antidromic studies. Orthodromic responses sometimes are low in amplitude, necessitating use of averaging techniques. Action potentials from distal muscles may obscure antidromic responses, because the thresholds of some motor axons are similar to those of large myelinated sensory axons. Fortunately, accurate measurement of SNAPs is still possible because the large-diameter sensory fibers conduct 5% to 10% faster than motor fibers. This relationship may change in disease states that selectively affect different fibers.

SNAPs may be obtained by several methods: (1) stimulating and recording a pure sensory nerve (such as the sural and radial sensory nerves); (2) stimulating a mixed nerve while recording distally over a cutaneous branch (such as the antidromic median and ulnar sensory responses), or (3) stimulating a distal cutaneous branch while recording over a proximal mixed nerve (such as in orthodromic median and ulnar sensory studies). Similar to their motor counterparts, sensory NCS record several measurements (Fig. 32B.2):

SNAP amplitude: This semiquantitatively measures the number of sensory axons that conduct between the stimulation and recording sites. The calculation is from the baseline to negative peak or from negative peak to positive peak, and expressed in microvolts. SNAP duration and area may be measured but are not useful because of significant temporal dispersion and phase cancellation (see later discussion).

Latencies: Sensory distal latencies are measured (in milliseconds) from the stimulus artifact to the peak of the negative phase (peak latency) or from the stimulus artifact to the onset of the SNAP (onset latency). A large shock artifact, a noisy background, or a wavy baseline may obscure onset latency. Although peak latency does not reflect the fastest-conducting sensory fibers, it is easily defined and more precise than onset latency.

Conduction velocity: This requires stimulation at a single site only because the latency consists of just the nerve conduction time from the stimulus point to the recording electrode. As with motor velocity, the calculation may be done using both distal and proximal stimulations. Only onset latencies (not peak latencies) are useful to calculate velocities to assess the speed of the fastest-conducting fibers.

Fig. 32B.2 Antidromic median sensory nerve conduction study after stimulation at the wrist, revealing peak and onset latencies and sensory nerve action potential amplitude, duration, and area. The shock artifact interferes with accurate determination of onset latency, whereas peak latency is easily determined.

Mixed Nerve Conduction Studies

Stimulating and recording from nerve trunks containing sensory and motor axons constitute mixed NCS. Often these tests require stimulating a nerve trunk distally and recording more proximally because large CMAPs contaminate the reverse by obscuring the lower-amplitude MNAPs. The MNAP may be of low amplitude or not elicitable when the nerve is deeply situated (as at the elbow or knee) because of tissue interposed between the nerve and the recording electrode. Therefore, MNAPs are limited to assessing mixed nerves in distal nerve segments in the hand or foot, such as the mixed palmar and mixed plantar studies used to evaluate carpal tunnel syndrome and tarsal tunnel syndrome, respectively.

Segmental Stimulation in Short Increments

Routine NCS are sufficient to localize the site of involvement in most patients with entrapment neuropathies. During evaluation of a focal demyelinating lesion, however, inclusion of the unaffected nerve segment in relatively long distal latency or conduction velocity calculation dilutes the effect of slowing at the injured site and decreases the sensitivity of the test. Therefore, incremental stimulation across a shorter nerve segment is useful to help localize an abnormality that might otherwise escape detection. Localization that is more precise entails “inching” the stimulus in short increments along the course of the nerve. The study of short segments provides better resolution of restricted lesions. For example, a nerve impulse may be found to conduct at a rate of 0.2 msec per 1.0 cm (50 m/sec). For a 1-cm segment, then, demyelination would double the conduction time to 0.4 msec/cm. In a 10-cm segment, normally covered in 2.0 msec, a 0.2-msec increase would constitute a 10% change, or approximately 1 standard deviation—well within the normal range of variability. However, the same 0.2-msec increase would represent a 100% change in latency if measured over a 1-cm segment. The large per-step increase in latency more than compensates for the inherent measurement error associated with stimulating multiple times in short increments.

The inching (or actually “centimetering”) technique is particularly useful in assessing nerve conduction in patients with carpal tunnel syndrome or an ulnar neuropathy at the elbow or wrist (McIntosh et al., 1998). For example, stimulation of a normal median nerve in 1-cm increments across the wrist results in latency changes of approximately 0.16 to 0.21 msec/cm from midpalm to distal forearm (Fig. 32B.3). A sharply localized latency increase across a 1-cm segment indicates a focal abnormality of the median nerve (Fig. 32B.4). An abrupt change in waveform usually accompanies the latency increase across the site of compression.

Fig. 32B.3 A, Twelve sites of stimulation in 1-cm increments along the length of the median nerve. The 0 level is at the distal crease of the wrist, corresponding to the origin of the transverse carpal ligament. Sensory nerve action potentials (SNAPs) and compound muscle action potentials are recorded from the second digit and abductor pollicis brevis, respectively. B, SNAPs in a normal subject recorded after stimulation of the median nerve at multiple points across the wrist. The site of each stimulus is indicated on the left. The latency changes increased linearly (approximately 0.16 to 0.21 msec) as the stimulus site was moved proximally in 1-cm increments.

(Reprinted with permission of the author and publisher from Kimura, J., 1979. The carpal tunnel syndrome: localization of conduction abnormalities within the distal segment of the median nerve. Brain 102, 619-635. By permission of Oxford University Press, Inc.)

Fig. 32B.4 Sensory nerve action potentials in a patient with bilateral carpal tunnel syndrome (see also Fig. 35B.3 for settings). A sharply localized slowing was found from point −2 to point −1 in both hands, with a latency change measuring 0.7 msec on the left (A) and 1.1 msec on the right (B), compared with the other segments with normal latency changes of approximately 0.16 to 0.21 msec. Note also a distinct change in waveform of the sensory potential at the point of localized conduction delay.

Physiological Variability and Common Sources of Error

The major pitfalls in NCS usually involve technical errors in the stimulating or recording system (Kimura, 1997). Common errors include large stimulus artifact, increased electrode noise, submaximal stimulation, co-stimulation of adjacent nerve not under study, eliciting an unwanted potential from distant muscles, recording or reference electrode misplacement, and errors in measurement of nerve length and conduction time. Other errors are attributable to intertrial and physiological variability, which include the effects of temperature, age, the length of studied nerve, anomalous innervation, and temporal dispersion.

Temperature

Nerve impulse propagation slows by 2.4 m/sec, or approximately 5%, per degree centigrade from 38°C to 29°C of body temperature. Also, cooling results in a higher CMAP and SNAP amplitude and longer response duration, probably because of accelerated and slowed sodium channel inactivation (Rutkove et al., 1997). Therefore, a CMAP or SNAP with high amplitude and slow distal latency or conduction velocity should raise the suspicion of a cool limb. To reduce this type of variability, a plate thermistor measures skin temperature. This measurement correlates linearly with the subcutaneous and intramuscular temperatures. If the skin temperature falls below 33°C, warm the limbs by immersion in warm water or by application of warming packs or a hydrocollator. Adding 5% of the calculated conduction velocity for each degree below 33°C theoretically normalizes the result. The use of such conversion factors is based on evidence obtained in healthy persons, however, and may not be applicable in patients with abnormal nerves.

Age

Because myelination is incomplete at birth, nerve conduction velocities are half the adult values in full-term newborns; in 23- to 24-week premature newborns, velocities are one-third the values for term newborns. They attain adult values at 3 to 5 years. Motor and sensory nerve conduction velocities tend to increase slightly in the arms and decrease in the legs during childhood up to the age of 19 years. Conduction velocities slowly decline after age 50, so the mean conduction velocity reduces by approximately 10% at 60 years of age.

Aging also diminishes SNAP and CMAP amplitudes, which decline slowly after age 60. SNAP amplitudes are affected more prominently, so much that normal upper limb SNAP amplitude drops to 50% by age 70, and lower limb SNAPs in many healthy persons older than 60 are low in amplitude or unevokable. Therefore, interpret absence of lower extremity SNAPs in older adults with caution; the finding is not necessarily abnormal without other confirmatory data.

Height and Nerve Segment Lengths

An inverse relationship between height and nerve conduction velocity suggests that longer nerves conduct slower than shorter nerves. For example, the nerve conduction velocities of the peroneal and tibial nerves are 7 to 10 m/sec slower than those of the median and ulnar nerves. The slightly lower temperature of the legs compared with the arms is not the entire explanation. Possible factors accounting for the length-related slowing include abrupt distal axonal tapering, progressive reduction in axonal diameter, and shorter internodal distances. For similar reasons, nerve impulses propagate faster in proximal than in distal nerve segments. Adjustments of normal values are necessary for patients of extreme height; this usually is no more than 2 m/sec below the lower limit of normal.

Anomalies

Several anomalous peripheral innervations may influence interpretation of the electrodiagnostic study. Two of these variants, the Martin-Gruber anastomosis and the accessory deep peroneal nerve, have a significant effect on NCS.

Martin-Gruber Anastomosis

In the Martin-Gruber anastomosis, anomalous fibers cross from the median to the ulnar nerve in the forearm. The communicating branches usually consist of motor axons supplying the ulnar innervated intrinsic hand muscles, particularly the first dorsal interosseous muscle, the hypothenar muscles (abductor digiti minimi), and the thenar muscles (adductor pollicis, deep head of flexor pollicis brevis), or a combination of these muscles. The Martin-Gruber anastomosis occurs in approximately 15% to 20% of the population and sometimes is bilateral. This anomaly manifests as a drop in the ulnar CMAP amplitude between distal and proximal stimulation sites (simulating the appearance of conduction block on ulnar NCS recording from the abductor digiti minimi or first dorsal interosseous). With distal stimulation (at the wrist), the CMAP reflects all ulnar motor fibers, whereas proximal stimulation activates only the uncrossed fibers, which are fewer in number. This anomaly can be confirmed by median nerve stimulation at the elbow, which evokes a small CMAP from the abductor digiti minimi or first dorsal interosseous, which is not present on median nerve stimulation at the wrist. When anomalous fibers innervate the thenar muscles, stimulation of the median nerve at the elbow activates the nerve and the crossing ulnar fibers, resulting in a large CMAP, often with an initial positivity caused by volume conduction of action potential from the ulnar thenar muscles to the median thenar muscles. By contrast, distal median nerve stimulation evokes a smaller thenar CMAP without the positive dip because the crossed fibers are not present at the wrist. In addition, the median nerve conduction velocity in the forearm is spuriously fast, particularly in the presence of carpal tunnel syndrome because the CMAP onset represents a different population of fibers at the wrist than at the elbow. Collision studies obtain an accurate conduction velocity by using action potentials of the crossed fibers (Sander et al., 1997).

Accessory Deep Peroneal Nerve

About 20% to 30% of subjects have an anomalous deep peroneal nerve. It is a branch of the superficial peroneal nerve and usually arises as a continuation of the muscular branch that innervates the peroneus longus and brevis muscles. It passes behind the lateral malleolus and terminates in the extensor digitorum brevis (EDB) on the dorsum of the foot. During peroneal motor NCS recording from the EDB, the peroneal CMAP amplitude is larger-stimulating proximally than distally because the anomalous fibers are not present at the ankle. Stimulation behind the lateral malleolus confirms this anomaly, which yields a small CMAP that approximately equals the difference between the CMAP amplitudes evoked with distal and proximal peroneal nerve stimulations. Complete innervation of the EDB by the accessory deep peroneal nerve is rare but should be suspected if there is preservation of function in the EDB muscle (i.e., extension of lateral toes) in a patient with severe deep peroneal neuropathy (Kayal and Katirji, 2009).

Pre- And Post-Fixed Brachial Plexus

In most people, the brachial plexus arises from the C5 to T1 cervical roots. In some, the plexus origin shifts one level up (pre-fixed), arising from C4 to C8, and in others, one level down (post-fixed), originating from C6 to T2. These anomalies result in error in the precise localization of cervical root lesions based on myotomal and dermatomal representation. In pre-fixed plexus, the location of the cervical lesion is one level higher than concluded from findings on the clinical examination and electrodiagnostic studies. In contrast, with post-fixed plexus, the cervical root lesion is one level lower.

Riche-Cannieu Anastomosis

Riche-Cannieu anastomosis is a communication in the palm between the recurrent motor branch of the median nerve and the deep branch of the ulnar nerve. The result is dual innervation of some intrinsic hand muscles such as the first dorsal interosseous, adductor pollicis, and abductor pollicis brevis. Riche-Cannieu anastomosis is rather common but often is not clinically or electrophysiologically apparent. When this anomaly is prominent, denervation in ulnar muscles may follow a median nerve lesion, and vice versa. In addition, a complete median or ulnar nerve lesion may be associated with relative sparing of some median innervated muscles or ulnar innervated muscles in the hand.

Temporal Dispersion and Phase Cancellation

The CMAP, evoked by supramaximal stimulation, represents the summation of all individual MUAPs directed to the muscle through the stimulated nerve. Typically, as the stimulus site moves proximally, the CMAP slightly drops in amplitude and area and increases in duration. The cause is temporal dispersion in which the velocity of impulses in slow-conducting fibers lags increasingly behind those of fast-conducting fibers as conduction distance increases. With dispersion, a slight positive and negative phase overlap, and cancellation of MUAP waveforms is seen (Fig. 32B.5). The result of temporal dispersion and phase cancellation is a reduction of CMAP amplitude and area and prolongation of its duration.

Fig. 32B.5 Compound muscle action potentials showing the relationship between fast-conducting (F) and slow-conducting (S) motor fibers. With distal stimulation (top), two unit discharges representing motor unit potentials sum to produce a muscle action potential twice as large. With proximal stimulation (bottom), motor unit potentials of long duration still superimpose nearly in phase despite the same latency shift of the slow motor fiber. Thus, a physiological temporal dispersion alters the size of the muscle action potential only minimally, if at all. Phase cancellation increases substantially when the latency difference between fast- and slow-conducting fibers is increased by a demyelinating neuropathy. This gives the false impression of motor conduction block.

(Reprinted with permission from Kimura, J., Machida, M., Ishida, T., et al., 1986. Relation between size of compound sensory or muscle action potentials and length of nerve segment. Neurology 36, 647-652.)

Physiological temporal dispersion affects the SNAP more than the CMAP (Fig. 32B.6). This difference relates to two factors. The first relates to the disparity between sensory fiber and motor fiber conduction velocities. The range of conduction velocities between the fastest and the slowest individual human myelinated sensory axons is almost twice that for the motor axons (25 m/sec and 12 m/sec, respectively). The second factor is the difference in duration of individual unit discharges between nerve and muscle. With short-duration biphasic sensory spikes, a slight latency difference could line up the positive peaks of the fast fibers with the negative peaks of the slow fibers and cancel both (Fig. 32B.7). In longer-duration MUAPs, the same latency shift would only partially superimpose peaks of opposite polarity, and phase cancellation would be less of a factor.

Fig. 32B.6 Simultaneous recordings of compound muscle action potentials (CMAPs) from the thenar eminence and sensory nerve action potentials (SNAPs) from index finger after stimulation of the median nerve at palm (P), wrist (W), elbow (E), and axilla (A). With progressively more proximal stimulation, CMAPs remained nearly the same; for SNAPs, however, both amplitude and the area under the waveform became much smaller.

Fig. 32B.7 Sensory nerve action potentials. A model for phase cancellation between fast-conducting (F) and slow-conducting (S) sensory fibers. With distal stimulation (top), two unit discharges sum in phase to produce a sensory action potential twice as large. With proximal stimulation (bottom), a delay of the slow fiber causes phase cancellation between the negative peak of the fast fiber and positive peak of the slow fiber, resulting in a 50% reduction in size of the summated response.

Intertrial Variability

Principal factors contributing to an intertrial variability include errors in determining surface distance and measuring latencies and amplitudes of the recorded response. Amplitudes vary most, probably reflecting a shift in the recording site. NCS are more reproducible when administered by the same examiner because a significant degree of interexaminer difference exists.

Electrodiagnosis by Nerve Conduction Studies

Although both NCS and needle EMG are required in most patients to confirm an electrodiagnostic impression, certain neuromuscular disorders may be evident on NCS alone.

Focal Nerve Lesions

Peripheral nerve is composed of unmyelinated and myelinated axons surrounded by Schwann cells and a supporting tissue. Surrounding the unmyelinated axons are only the plasma membranes of Schwann cells. By contrast, wrapped around myelinated axons are multiple myelin layers that have a low capacitance and large resistance. Surrounding the myelinated axon is myelin, along with Schwann cells, except at certain gaps called the nodes of Ranvier, where sodium channels are highly concentrated and saltatory conduction occurs. Three supportive layers—the endoneurium, perineurium, and epineurium—are highly elastic and protect the myelin and axon from external pressure and tension surround nerves. Nerve fibers may be injured by a variety of mechanisms including compression, ischemia, traction, and laceration.

The classification of peripheral nerve lesions follows. In neurapraxia (first-degree injury), distortion of myelin occurs near the nodes of Ranvier, producing segmental conduction block without wallerian degeneration. In axonotmesis (second-degree injury), the axon is interrupted, but all the supporting nerve structures remain intact. In neurotmesis, the nerve injury is severe, resulting in complete disruption of the nerve with all the supporting structures (see Chapter 52D). Often, the neurotmesis category is divisible as follows: third-degree injury, with disruption of the endoneurium and with intact perineurium and epineurium; fourth-degree injury, with disruption of all neural elements except the epineurium; and fifth-degree nerve injury, with complete nerve transection resulting in complete discontinuity of the nerve. Electrodiagnostic studies alone cannot accurately distinguish among the five degrees of nerve injuries, but they can separate the first (neurapraxia) from the other types (Wilbourn, 2002).

Demyelinative Mononeuropathy

When focal injury to myelin occurs, conduction along the affected nerve fibers may alter. This may result in conduction slowing or block along the nerve fibers. The cause of conduction block is interruption of action potential transmission across the nerve lesion; it is the electrophysiological correlate of neurapraxia and usually results from loss of more than one myelin segment (segmental or internodal demyelination). Bracketing two stimulation points, one distal and one proximal to the site of injury, best localizes a nerve lesion with conduction block. With such lesions, stimulation distal to the lesion elicits a normal CMAP, whereas proximal stimulation evokes a response with reduced amplitude or fails to evoke any response, defined as partial or complete conduction block, respectively (Fig. 32B.8, A). Several limitations exist to the diagnosis of demyelinative conduction block:

1. Phase cancellation between peaks of opposite polarity may reduce CMAP size because of abnormally increased temporal dispersion. Such excessive desynchronization often develops in acquired demyelinative neuropathies. If the distal and the proximal responses have dissimilar waveforms, the discrepancy in amplitude or area between the two may be the result of phase cancellation rather than conduction block. Therefore, for a diagnosis of partial conduction block, findings should include a significantly lower CMAP amplitude and smaller CMAP area with stimulation proximal to the injury site than with the CMAP distal to it, and without any significant prolongation of CMAP duration. More than 50% decay of both the CMAP amplitude and area across the lesion usually is the criterion for definite conduction block.

2. Distal demyelinating lesions causing conduction block of the nerve segment between the most distal stimulating point and the recording site manifest as unelicitable or low CMAP amplitudes at both distal and proximal stimulation sites. This finding mimics the NCS seen with axonal degeneration.

3. The prominent temporal dispersion normally seen in evaluating SNAPs precludes the use of these potentials to diagnose conduction block.

4. Conduction block also may follow axonal loss before the completion of wallerian degeneration.

Fig. 32B.8 Findings on nerve conduction studies. A, Demyelinative conduction block. Note the proximal compound muscle action potential (CMAP) is either low in amplitude (partial block) or absent (complete block). B, Focal synchronized slowing of the distal nerve segment (a) or the proximal nerve segment (b). C, Focal desynchronized slowing of the forearm nerve segment, resulting in significant dispersion of the proximal CMAP. D, Axon loss (partial), studied early and late after nerve trauma. R, Recording; S, stimulation.

(Reprinted with permission from Wilbourn, A.J., 2002. Nerve conduction studies. Types, components, abnormalities and value in localization. Neurol Clin 20, 305-338.)

Focal slowing of conduction usually is the result of widening of the nodes of Ranvier (paranodal demyelination). Slowing, often synchronized, affects all large myelinated fibers equally. This results in prolongation of distal latency if the focal lesion is distal (see Fig. 32B.8, B,a), or slowing in conduction velocity if the focal lesion is proximal (see Fig. 32B.8, B,b). CMAP amplitude, duration, and area, however, are normal and do not change when the nerve is stimulated proximal to the lesion. Desynchronized slowing (differential slowing) occurs when conduction velocity reduces at the lesion site along a variable number of the medium or small nerve fibers (average- or slower-conducting axons). Here, the CMAP disperses with prolonged duration on stimulations proximal to the lesion. The speed of conduction along the injury site (latency or conduction velocity) is normal because of sparing of at least some of the fastest-conducting axons (see Fig. 32B.8, C). When synchronized and desynchronized slowing coexists, slowing of distal latency or conduction velocity accompanies the dispersed CMAP with prolonged duration.

Axon-Loss Mononeuropathy

After acute focal axonal damage, the distal nerve segment undergoes wallerian degeneration. Characteristically, unelicitable or low CMAP amplitudes with distal and proximal stimulations are signs of complete or partial motor axonal loss lesions. The CMAP amplitudes provide a reliable estimate of the amount of axonal loss except in the chronic phase, in which reinnervation via collateral sprouting often increases the CMAP and gives a misleadingly low estimate of the extent of original axonal loss.

In partial axon-loss lesions, distal latencies and conduction velocities are normal or borderline. Selective loss of fast-conducting fibers associated with more than a 50% reduction in mean CMAP amplitude can slow conduction velocity up to 80% of normal value because the velocity represents the remaining slow-conducting fibers. Motor conduction velocity may slow to 70% of normal value with reduction of CMAP amplitude to less than 10% of the lower limit of normal.

Soon after axonal transection (i.e., for the first 48 hours), the distal axon remains excitable. Therefore, stimulation distal to the lesion elicits a normal CMAP, whereas proximal stimulation elicits a response with reduced amplitude, producing a conduction block pattern (see Fig. 32B.8, D, middle panel). This pattern is axonal noncontinuity, early axon loss, or axon-discontinuity conduction block. Soon, however, the distal axons undergo wallerian degeneration, and the distal CMAP decreases to equal the proximal CMAP (see Fig. 32B.8, D, lower panel). With wallerian degeneration, the distal CMAP decreases in amplitude starting 1 or 2 days after nerve injury and reaches its nadir in 5 to 6 days. In contrast, the distal SNAP lags slightly behind and reaches its nadir in 10 or 11 days (Fig. 32B.9). The difference between the decline of the SNAP and CMAP amplitudes after axon loss probably relates to neuromuscular transmission failure, which affects only the CMAP amplitude. Supporting this hypothesis is the fact that MNAPs recorded directly from nerve trunks follow the time course of SNAPs.