Unraveling the mysteries of nerve conduction testing
With recent advances in serologic testing and molecular diagnostics, our understanding of the peripheral nervous system has grown ever more sophisticated. But the basics of nerve conduction studies, or the electrodiagnostic testing we do on peripheral nerves, this really hasn’t changed much over the last century. We’re still doing this a lot like it’s always been done, with very few exceptions. So if you’d like to know what this is that neurologists do in the EMG lab, then you’ll really like our show this week. Today on BrainWaves, we’re going to cover a few of the basic concepts that underlie nerve conduction testing, with a particular emphasis on conduction block.
It all starts with a zap. And to know how the nerves transmit the information from that electrical impulse, you have a stimulating site and a recording site. Using the ulnar sensory nerve as an example, the technician will stimulate the pinky finger and record from the medial wrist for an orthodromic response. (Orthodromic just means that the stimulation and recording take place in the direction of the natural signal. Antidromic is the opposite.) Once you get a response from your recording electrode, then you have a lot of information to work with, specifically:
- How large is that response?
- How quick was that response?
The SIZE OF THE RESPONSE is only measured by one value, the amplitude. When it’s the sensory nerve action potential, it’s the SNAP—and when it’s the motor equivalent, it’s called the compound motor action potential—or CMAP. Second, the speed of conduction is measured in a few ways, and each way gives you different data on where the nerve could be injured or how it could be injured. First, you get a peak latency. This is the time it takes for the stimulus to reach the peak of the first negative deflection of the recording electrode. Then you get a SNAP or CMAP duration, which is the time it takes from the onset of the recorded potential to the first baseline crossing. When you’re looking at it on the screen, it’s the time it takes for the hump to start and finish. The duration tells you the time difference between when the first and the last nerve signals arrive at the recording site. Lastly, there is conduction velocity. In sensory nerves, this is calculated as the distance between the stimulating and recording sites divided by the time it takes for the stimulus to reach the recording electrode.
Now I said motor responses are a little more complicated. Here’s what I meant by that. While you only need a single stimulation in order to calculate a sensory nerve conduction latency, you need two stimulations to calculate a motor latency (see above figure). That’s because the distal latency of a motor study, or the time it takes for the stimulation to result in a muscle response, is a reflection of several physiologic events. The distal motor latency captures of the nerve conduction time from the distal stimulating site to the neuromuscular junction (A), the neuromuscular junction transmission time (B), and the muscle depolarization time (C). Let’s presume your distal stimulating site in this example is near the wrist for the median nerve. Then you take a more proximal stimulation, say at the elbow, and record the latency from that signal to result in muscular depolarization. Now you have the proximal latency from the elbow to the abductor pollicis brevis depolarization and the distal latency from the wrist to the APB. By subtracting the distal latency from the proximal latency, you can get a pure latency of the motor nerve across that distance between the two points of stimulation.
Here’s an analogy. Imagine you wanted to measure the time it takes for LeBron James to sprint across the court to make a slam dunk. What is the velocity of LeBron during the sprint? See the figure below. You can measure the time it takes him to run from one far end of the court to another—the proximal latency—but as he gets nearer to the basket, at about the free throw line, he is going to slow down in order to control the ball and make the basket (A). And once he reaches the basket, it’s only his momentum that continues to push him forward to the end of the court from there (C + B). He is not sprinting anymore. So measuring the latency from one end of the court to the other isn’t going to help you accurately calculate the velocity of his sprint. You want to capture the latency between two points where he is deep in stride and subtract out the rest. So you measure the latency from the far end of the court to the other—the proximal latency. And you measure the latency from the free throw line to the end of the court—the distal latency (A + B + C). Then you subtract that free throw latency from the total latency, and you measure that distance, and now you can calculate the velocity of his sprint before he has to slow down at the free throw line.
Now what do you do with this information once you have it? What do all the numbers mean and how do you interpret them? As a general principle, when evaluating for nerve damage, you’ve got to always be thinking bigger picture:
- Is peripheral nerve damage present, or is it absent?
- Is the nerve damage consistent with axon loss or demyelination?
- Is it distally predominant or is it diffuse?
- Is it unifocal or multifocal?
While you’ve got to start the investigation by looking at each of these individual numbers describing speed and amplitude, you’re also trying to tease out these features which are going to be useful in making a clinical diagnosis. Start with the sensory nerve responses. Because if sensory nerves are normal, then that means there’s either (A) nothing discernibly wrong with the peripheral nerves, (B) something’s wrong but it’s not affecting the large sensory fibers—like a purely motor problem, or a small fiber neuropathy, or (C) something’s wrong and it’s proximal to the dorsal root ganglion. This last point is a major pearl. Because the dorsal root ganglion lies outside the neural foramen, foramenal compression will not injure the DRG or the distal nerve. So with a radiculopathy for example, the patient may have a sensory disturbance, but the sensory responses will be normal because the DRG and its afferent axon are intact.
Case No. 1
Starting right to left, we’ll begin with the SNAP. The left sural nerve SNAP is 2.9 uV, which is small—it should be at least 6 in an adult—and the left peroneal SNAP is “NR,” which means no response. So we have identified a lesion in these afferent sensory nerves, and the lesion is distal to the dorsal root ganglion. But is the nerve damage to the axon or to the myelin? Going back to the basics, the SNAP amplitude, like the CMAP amplitude, reflects the number of underlying axons. So when you have axon loss, your SNAPs and CMAPs will fall. And some patients may just have low amplitudes, or it may be technician-dependent, so when you can, comparing the same nerve on each side can be helpful to determine if one SNAP is pathologically low or if its not. In this case, we don’t have a contralateral leg study. So all we can say is that the sural and peroneal SNAPs are small—and this means axon loss, right?
No! At least not exactly. With axon loss, you’re correct to think that one of the first changes you’ll see in nerve conduction testing will be a drop in the SNAPs or CMAPs. But a small SNAP doesn’t necessarily indicate axon loss. A small SNAP can also be consistent with demyelination, as can a small CMAP. And when you see the amplitude of a CMAP fall by more than 50% between proximal and distal stimulation sites, this is called conduction block. Or—more rarely—small CMAPs can be seen in severe myopathy because CMAPs are also dependent on neuromuscular junction transmission and the motor response to the stimulation. Here is one of the main take away points for this entire episode: SNAPs and CMAPs are not totally useful in distinguishing axon loss from demyelination. In acquired demyelinating neuropathies, where the demyelination is patchy, there is slowed saltatory conduction along some parts of some of the peripheral nerves. So if you take a bulky peripheral nerve, some of the individual nerve fibers are going to be signaling normally and quickly—like 40 to 60 cm/s—and some are going to be signaling slowly—maybe 20-30 cm/s. Therefore, when you are recording at a distal site, the summated response is going to be lower than what is anticipated because all the signals are arriving at separate times. Remember when I talked about duration earlier? How duration is the time it takes from when the first signal and the last signal reach the recording site? A longer duration indicates there is a larger difference in the velocity of individual fibers. And electrographically, when we visualize this nerve signal, this is called temporal dispersion. So again, when you see these features, conduction block and temporal dispersion, they are all going to be features of an acquired pattern of demyelination like AIDP, multifocal motor neuropathy, or the POEMS syndrome. In congenital cases of demyelination like in CMT1a, due to a mutation in PMP-22, you won’t see these characteristics where some nerves are partially slowed and others that are normal. All nerves are going to be equally slowed. But back to our example here, case no. 1. Looking at the raw numbers themselves, seeing that the SNAP is low in the sural nerve, we’ll need more information to decide where the pathology lies: the axon or the myelin sheath.
To determine if demyelination is present, we need evidence that there is slowed saltatory conduction. As we mentioned before, this is quantified by conduction velocity, distal latency, and the duration. Starting with the conduction velocity, which I think is the easiest to follow, in case 1 the conduction velocity of that left sural nerve is 40 cm/s. A good rule of thumb is that velocities in the legs should be about 40, and in the arms it should be about 50. So this sural nerve velocity of 40 what we expect it to be, about 40. Looking at the left peroneal sensory nerve, the conduction velocity is zero. This is because there was no sensory nerve response recorded. So a velocity can’t be calculated. Together, this is not a great picture for demyelination. It seems more axonal. But before we reach that conclusion, let’s take a look at the rest of the nerves and see what they show.
In the left arm, the radial nerve has an amplitude of 6.5, much lower than the normal of 15-20 or so, but the velocity is normal. So it’s also got an axonal injury pattern. The left median nerve amplitude should be around 15-20, but the amplitude between digit II and the wrist is only 5.7, with a conduction velocity of 49—close enough to 50. Another axonal pattern. The ulnar sensory amplitude should be about 10, and one of the readings is a 6, so a little lower than expected. But the conduction velocity is normal in the 50s. Overall, this is a pattern of low sensory amplitudes without slowing. Meaning axon loss. And if I had to say if it was focal or multifocal, distally predominant or diffuse, I would say it appears pretty diffuse, involving the left upper and lower extremity sensory nerves. We’ll skip over the interpretation of the motor nerves for now, but I’ll just acknowledge a similar pattern was observed in the CMAPs throughout the left side, and there was no evidence of demyelination. So this study was read as a moderate-to-severe sensorimotor axonal polyneuropathy. Disorders consistent with this would be severe diabetes, uremia, and some nutritional deficiencies.
Case No. 2
Starting with the sensory nerve responses—again to determine if there is a lesion present, and if that lesion is proximal or distal to the dorsal root ganglion—looking at the left radial, median, and ulnar nerves, there is no response, NR. So the lesion is distal to the dorsal root ganglion, but we don’t know if it’s demyelinating or axonal, distally predominant or diffuse.
Looking at the motor studies next, the median motor amplitude, the median CMAP, drops from 2.9mV to 1.4mV as you record from proximal to distal. A normal median CMAP should be more than 4mV, so this CMAP is already lower than it should be, at 2.9. Then it falls even further to 1.4mV with distal stimulation. This is your example of conduction block, a fall in the CMAP of more than 50% between proximal and distal stimulation. But this doesn’t necessarily rule out severe axon loss, so we need to know if it’s related to any slowing of the conduction velocity or prolonged distal motor latency. Moving right to left on our chart, the conduction velocity is 17. 17 cm/s when it should be 50 for the median nerve—remember, 50 for the arms, 40 for the legs. This velocity of 17 is consistent with demyelination, which we define as a conduction velocity slowing to 70% or less of what is expected. In the arms, knowing the conduction velocity should be 50, anything that’s less than 35 would be consistent with demyelination. And in the legs, where conduction velocities should be 40, a velocity less than 28 would be concerning for myelin injury. Now, let it be known that severe axon loss can also cause some mild slowing, but it’s usually not to 70% of what’s expected. Slowing due to severe axonal loss is going to be between 35 and 45 cm/s in the arms and 28-35 cm/s in the legs.
For Case no. 2, the study is consistent, at least for now, with demyelination and conduction block of the left median motor nerve. Often this indicates one of the acquired demyelinating processes like we mentioned before: AIDP, CIDP, MMN, or POEMS. But let’s look through the rest of the nerves before we jump to conclusions. The left ulnar CMAP is lower than expected, 2.6 when it should be at least 4, but it does not fall below 50% from proximal to distal stimulation. So not in the conduction block range, but still low. The velocity here is also quite slow, 15 to 22 cm/s when it should be 50. So, now there is evidence of ulnar demyelination, but without conduction block. For this patient, the phrenic nerves were also tested, because the patient had an elevated hemidiaphragm on chest x-ray. And it does appear that the phrenic nerves also have very small CMAPs. Altogether, a pattern consistent with diffuse demyelinating disease and possibly conduction block. Probably an acquired demyelinating disorder.
But what if I told you that the motor nerve duration was normal, meaning there was no temporal dispersion on visualization of the CMAPs? The CMAP morphology in this patient’s nerves was normal, just a lot smaller in amplitude than expected. This indicates that each of the individual motor nerve fibers were signaling at the same slow velocity and reaching the recording site at nearly the same time, suggesting that whatever the demyelinating event was to begin with, it started in each of the nerves at the same time and progressed at the same rate. Only a heritable form of demyelination can cause this, and that’s because the genes responsible for myelin production are equally dysfunctional in all of the peripheral nerves. It turns out this patient had a mutation in PMP22, which is causative for CMT1a. And sometimes you can get conduction block in Charcot Marie Tooth, but you shouldn’t see temporal dispersion or phase cancellation. So that’s what we found in her nerve conduction study.
Wrapping up. I think this is a good start to the fundamentals of nerve conduction testing. We opened the show kind of talking about the basics of nerve conduction studies and the basic principles of interpreting these studies—identifying axonal patterns from demyelinating patterns, the importance of distinguishing distal lesions from diffuse lesions, and unifocal versus multifocal processes. Always be looking at the bigger picture. Now, we didn’t cover everything this week, and there was nowhere near enough time to go over how to perform a nerve conduction study. For example, we didn’t even acknowledge the ground electrode, which provides your zero voltage reference point. We didn’t talk about when nerves begin to show signs of injury on conduction testing—which is why a patient with severe GBS can look like their nerves are totally normal during the first week of symptoms. We also didn’t get into all the exceptions to the rules of nerve conduction testing. For example, I told you that conduction block is associated with acquired demyelinating processes, like AIDP, but it is also seen in some cases of axonal neuropathies like AMAN—acute motor axonal neuropathy. And we didn’t even get to the EMG, which is a complementary study to the nerve conduction test. I guess that will all have to wait for another time. Until then, happy holidays.
The BrainWaves podcast and online content are intended for medical education and entertainment purposes only. Thank you for reading our disclaimer. And the extremely long article this week. Now go have a nice day.
- Allen JA. Chronic Demyelinating Polyneuropathies. Continuum (Minneap Minn). 2017;23:1310-1331.
- Kincaid JC. Neurophysiologic Studies in the Evaluation of Polyneuropathy. Continuum (Minneap Minn). 2017;23:1263-1275.
- Preston DC and Shapiro BE. Electromyography and neuromuscular disorders : clinical-electrophysiologic correlations. 3rd ed. London ; New York: Elsevier Saunders; 2013.
- Pareyson D, Scaioli V and Laura M. Clinical and electrophysiological aspects of Charcot-Marie-Tooth disease. Neuromolecular Med. 2006;8:3-22.