Motor nerve conductions are performed by stimulating a motor nerve at two separate locations along its course and recording from a surface electrode placed over the belly of a muscle supplied by that nerve. The amplitude of the evoked compound muscle action potential (CMAP) , distal latency (time from the distal stimulation to the onset of the corresponding motor response), and conduction velocity between the two stimulation sites are then calculated (Fig. 7.1). Based on comparisons to normative data, these values are then used to determine if the motor NCS is normal or shows evidence of a demyelinating or axonal lesion. As a general rule, motor NCS in axonal neuropathies (with intact myelination) exhibit reduced CMAP amplitudes with preserved conduction velocities and distal latencies, whereas demyelinating neuropathies feature slow nerve conductions, prolonged distal latencies, and normal or relatively normal CMAP amplitudes. Focal demyelination can also block conduction in some (partial) or all (complete) axons at the affected site despite the presence of structurally preserved axons. Similar to axon loss, motor conduction block produces clinical weakness. In a motor NCS performed on a nerve affected by a partial demyelinative conduction block, the CMAP evoked by stimulation above the site of block has significantly smaller amplitude than that evoked by stimulation below the block (Fig. 7.2). If the block is complete, no CMAP is evoked with proximal stimulation. A partial motor conduction block is the hallmark of an acquired demyelinating neuropathy.

A suprathreshold motor nerve stimulation evokes both an orthodromically (proximal to distal) and antidromically (distal to proximal) conducted neural impulse. The routine motor NCS depends on the orthodromic response. The antidromic impulse depolarizes the anterior horn cells of the constituent motor axons. If one or more of these cells is depolarized to or above its threshold, a recurrent discharge is generated that travels back down the nerve orthodromically resulting in “late” reactivation of the muscle, the F-wave response . In a single stimulation, only a small subset of motor neurons generates an F-wave. With multiple stimulations, consecutive F-waves are usually generated by different motor neurons, resulting in F-waves of varying latency and morphology (Fig. 7.3). If, on the other hand, the large Ia afferent sensory fibers are preferentially activated, typically by a lower intensity, longer duration stimulus, the evoked antidromically conducted impulse will synapse with and depolarize a pool of anterior horn cells in the spinal cord. These monosynaptic connections result in reflex activation of an orthodromically conducted motor impulse, which is recorded from the innervated muscle as the H-reflex. Unlike an F-wave , the latency of an H-reflex does not vary between stimulations, but its amplitude increases with progressively higher stimulus intensities until reaching a maximum. Then, with still higher stimulus intensities, the H-reflex decreases in amplitude and eventually disappears (Fig. 7.4). Whereas F-waves can be generated from all motor nerves and recorded from all muscles, H-reflexes are only consistently recorded from the soleus muscle after stimulation of the tibial nerve and more inconsistently recorded from the flexor carpi radialis muscle after stimulation of the median nerve.

Clinically, F-waves and H-reflexes are used to evaluate the proximal portions of peripheral nerves not amenable to standard motor NCS. Whereas, prolonged F-wave and H-reflex latencies generally occur in demyelinating lesions, absent F-waves and H-reflexes can result from both axonal and severe demyelinating insults. F-waves that “repeat,” i.e., recur with the same latency and waveform morphology suggest a dropout of the original number of anterior horn cells or their motor axons. Focal or multifocal demyelinating lesions frequently produce variable effects on different nerve fibers in the same motor nerve, resulting in an increased spectrum of conduction velocities within individual motor fibers and hence an increased range of F-wave latencies.

A sensory nerve conduction study is performed by stimulating a sensory nerve at one or more sites along its course and recording from a surface electrode placed over the same nerve distal or proximal to the stimulation site. Sensory conductions can be performed either antidromically or orthodromically. The evoked sensory nerve action potential (SNAP) is then evaluated with respect to amplitude, onset latency (time from stimulus to onset of the evoked potential), peak latency (time from stimulus to peak amplitude of the evoked potential), and conduction velocity (Fig. 7.5).

Sensory responses are generally much smaller than motor responses. As such, sensory NCS are more technically challenging and produce more variable results than motor NCS. Similar to motor studies, SNAPs in axonal neuropathies are reduced in amplitude or absent with normal or only mildly abnormal latencies and conduction velocities. Sensory responses can be slowed in demyelinating neuropathies, but they more commonly lose amplitude because of the increased range of individual nerve fiber conduction velocities. This widened spectrum of velocities results in decreased synchrony and increased phase cancellation of the waves generated by single fibers at the site of the recording electrode. Peripherally directed sensory nerve axons originate from cell bodies of primary sensory neurons located in the dorsal root ganglia. They only degenerate with corresponding loss of SNAP amplitude when lesions damage the dorsal root ganglion or the sensory axons themselves, i.e., neuropathies or plexopathies. On the other hand, peripheral sensory nerves and thus SNAPs are preserved in most radiculopathies, in which centrally directed sensory axons in the dorsal root proximal to the dorsal root ganglion are primarily affected.

There are some limitations to NCS. First, the nerve must be readily accessible to superficial stimulation with an equally accessible recording zone. This precludes NCS of most proximal nerves and plexus structures. Spinal nerve roots can be stimulated percutaneously with a needle electrode, but root stimulation is painful, technically difficult, and generally unreliable. Second, NCS only evaluate large sensory and motor fibers, so any process that is restricted to small-diameter nerve fibers may remain undetected with standard NCS. Third, sensory conductions can be technically difficult because of the inherently low amplitude of the evoked responses, especially in patients with extra tissue between the recording site and nerve (obesity and peripheral edema). Their performance requires utmost attention to proper technique to ensure the presence of a reproducible response. Undesired motor responses are not infrequently mistakenly identified as sensory responses when the SNAP is truly absent. Fourth, small cutaneous nerves—whether proximal or distal—cannot be reliably studied due to their inherently small surface recorded potentials (e.g., medial femoral cutaneous nerve). Fifth, the electrodiagnostic consultant needs to be attuned to the possibility of anomalous innervation pathways, e.g., Martin–Gruber (median-to-ulnar) anastomosis in the forearm, accessory peroneal branch in the lower leg, and aberrant course of the superficial peroneal sensory nerve. Sixth, CMAP amplitudes are exquisitely dependent on proper placement of the recording electrode over the “motor point” of the muscle, since the amplitude can change significantly with even minor relocations of this electrode. Seventh, preexisting conditions always need to be considered, since preexisting polyneuropathies and nerve entrapments are common and may confound interpretation of a postinjury NCS.

Another salient limitation is the delayed onset of NCS changes in response to acute axon-loss lesions. In traumatic nerve lesions resulting in Wallerian degeneration of the transected axons, which includes most needle injuries, the portion of the axon distal to the injury remains structurally intact and capable of transmitting a distally directed nerve impulse for the next 2–3 days prior to undergoing Wallerian degeneration (which will last up to 1 week). As such, CMAP and SNAP amplitudes do not begin to decrease until 2–3 days after the injury. CMAP amplitudes do not reach their nadir until 9 days after the insult. Sensory nerves degenerate at a slightly slower rate than motor nerves; accordingly, SNAP amplitudes do not achieve their nadir until 10–11 days postinjury [15]. Hence, NCS performed immediately after the injury serve as a preinjury baseline against which a follow-up study performed after 12 days can be compared to capture the true injury-related effects on the NCS.

There are also several practical limitations to the clinical utility of late responses. F-waves and H-reflexes are often normal in incomplete proximal axon-loss lesions due to the presence of many unaffected motor and sensory fibers that are capable of generating a normal response. These late responses are also often normal in proximal demyelinating lesions wherein focal slowing is limited to a small segment of the entire nerve pathway or conduction is preserved through the affected region by nondemyelinated nerve fibers. Moreover, F-waves are typically mediated by two or more nerve roots. As such, they tend to be normal in axonal or demyelinating mono-radiculopathies in which adjacent roots are spared. Unlike F-waves, H-reflexes are restricted in distribution and can only be recorded from the soleus (and occasionally the flexor carpi radialis) muscle in adults. H-reflexes tend to disappear in elderly individuals without known focal neurologic injuries.

For EMG , a needle electrode is inserted into a muscle for the purpose of directly recording the electrical activity (voltage changes) of the muscle fibers. In current EMG practice, two types of recording needle electrode are used, the concentric and the monopolar . A concentric needle consists of a cannula (reference electrode) and a core (active electrode), which are generally made of different materials. A monopolar needle consists of the active electrode alone covered with an insulating material except for its sharpened tip, which is the conductive area of the needle. When employing a monopolar recording electrode, the reference electrode is a surface electrode placed on the skin in close proximity to the active monopolar needle.

There are three separate phases of the evaluation. First, the electrical activity is evaluated during insertion of the needle electrode (i.e., “insertional activity”). In the second phase, the consultant searches for abnormal “spontaneous activity ” within the muscle which occurs when the electrode is stationary in the resting muscle. Multiple sites are assessed in each examined muscle. The type of abnormal spontaneous activity of greatest relevance to regional anesthesia-related nerve injury is the so-called fibrillation potential . A fibrillation potential is an EMG waveform generated by spontaneous firing of a single muscle fiber. Most fibrillation potentials discharge at a regular rate (Fig. 7.6). Positive sharp waves have the same origin and significance as fibrillation potentials. Fibrillations develop in acutely denervated muscle fibers, which become supersensitive after a short delay following the injury. They also occur in many muscle diseases associated with segmental necrosis of muscle fibers or primary electrical instability of the muscle fiber membrane.