What I want to do here is try to help with understanding of EMG reports. While there may be some need to explain some technical information about electromyography and its terms, this will be with the idea that this is needed to understand how to use the results.
Parts of an EMG There is some understandable confusion about this, since the test as a whole is called an "EMG", yet the individual parts are "NCS" or Nerve Conduction Studies, and "EMG", meaning needle electromyography. Most patients should have both parts, since they do not substitute for each other but are complementary.
Although a referring physician will indicate what should be done, typically in terms of an area or areas of the body, it is up to the electromyographer to decide precisely which nerves and muscles to test.
Nerve conduction studies can be used in 3 different ways:
A nerve may be a "nerve of interest", for example median nerve studies where carpal tunnel syndrome is suspected.
A nerve may be a useful comparison, for example ulnar nerve studies where carpal tunnel syndrome is suspected, or the opposite median nerve in an asymptomatic hand.
A nerve may be tested as a sense of general nerve health, eg, some diffuse polyneuropathy, such as we might do when we suspect diabetic polyneuropathy, or some kind of demyelinating polyneuropathy. Even when the clinical syndrome suggests a focal neuropathy, there is value in seeing whether there is some general nerve sensitivity to pressure or whether this is clearly just a focal phenomenon.
The decision is based on a combination of the clinical considerations raised by the patient's symptoms and exam findings, and sometimes other tests, such as an MRI scan which might suggest a pinched nerve root somewhere. Finally, the ongoing results may suggest some possibilities initially not apparent which need further investigation, or that some initial possibility is not worth pursuing.
Nerve Conduction Studies
There are quite a number of nerves which can be evaluated, but in general there is a relatively short list of nerves commonly tested, mainly related to accessibility, but also clinical usefulness. For example, there are common compressive syndromes involving the median nerve at the wrist, ulnar nerve at the elbow, and peroneal nerve at the fibular head, so by far these nerves are more commonly studied. Most nerves tested are mixed nerves, containing both motor and sensory nerve fibers, but where the nerve is stimulated and where the signal is subsequently picked up determine whether we call this a motor nerve conduction study or a sensory nerve conduction study. Let's use median nerve studies as an example.
The median nerve has a very dependable anatomical location, running down from the axilla, deep in the upper arm, then coming close to the surface in the antecubital region, then going deep again in the forearm, finally coming closer to the surface again near the carpal tunnel, then on into the hand. It is a mixed nerve all the way down to the hand, where it divides into motor branches mostly traveling to the thenar eminence, and sensory branches to the digits on the radial aspect of the hand. This splitting up allows us to test either motor or sensory function.
Nerve conduction studies can be done stimulating a nerve then picking up a signal either proximal or distal to the site. My preferred method for testing median sensory conduction is to stimulate the digital nerves, usually of the index finger, then recording the signal at the wrist. Many other electromyographers will stimulate at the wrist, then record at the digital nerves, but the principle is the same – if we record from or stimulate a pure sensory nerve, we will only study the sensory nerve conduction regardless of whether the other end of the process is a mixed nerve or not.
NCV or No NCV? My preference is to calculate an NCV or nerve conduction velocity whenever I can. The equipment (EMG machine) will allow us to measure a latency or the time (in milliseconds) it takes for a signal to travel from one place to another. By measuring the distance from stimulation point to recording point, we can then calculate the NCV for that interval, expressed in meters/second. The other method, probably more commonly used, is to use a fixed distance between stimulation and recording, then simply express the result as the latency.
An Idealized Sensory Waveform
S = Stimulus point, T = Takeoff point, P = Peak
The time (latency) from S to T is typically about 3 milliseconds.
The amplitude would be measured in microvolts (μV).
Although it would seem that you could on your own then calculate an NCV, the two techniques are not quite the same. The standard for calculating NCVs is to measure the latency of the fastest conducting fibers, which are represented by the very beginning of the waveform, also know as the take-off point, and this is where I take my latency measurements for my NCV calculation. Electromyographers who use this fixed distance technique typically take their latency measurement from the peak of the waveform, not really the fastest part of the curve, so this would not be a comparable NCV. There is ample literature standardizing the range of normal latencies of this peak, but it is important to stay with the latency and not calculate NCVs. My view is that this is a less precise method, which is why I continue to use my method, and in addition, with modern digital equipment, sensory nerve waveforms and their measurements are much easier than in the past.
Normal Values The bottom line is that you rely on your electromyographer to say whether a result is normal or not. Generally, for my method, a median sensory NCV should be 41 m/sec or greater, though other factors such as age of the patient and temperature of the hand may affect this assessment. There is also some information to be gained from the size of the response (amplitude of the waveform) but the range of normal is very great and consequently this is less helpful.
The Problem With Numbers There is a natural tendency in science and medicine to put emphasis on numbers. Numbers lead to the statistical analyses by which we say whether something is normal or abnormal, significant or insignificant. But we also have to remember that when we are talking about many measurable functions, like NCVs, we know that the normal population will have a Bell-shaped curve, and some overlap between normal and disease states exists. For example, it doesn't really make sense to think that an NCV of 41.0 m/sec is clearly normal, and 40.9 m/sec is clearly abnormal. So it's important to keep this in mind as we look at nerve conduction studies, and also helps to understand that the results of an EMG do not dictate treatment in any absolute way – other factors must always be considered.
What If You Get No Waveform? Even using signal averaging techniques, sometimes no discernable waveform can be seen, which I typically report as, "No response despite averaging". The best equipment may still have technical problems. One way of suggesting this is a valid response (valid meaning that the nerve is so sick that any response is too small to record) is to check another sensory nerve, and if necessary, then go back and verify that the original nerve still has no response. The problem with an absent response is that it is nonlocalizing. It tells you that electrophysiologically this nerve is "gone", but doesn't indicate where the problem is. For example, an absent median sensory response could be due to severe carpal tunnel syndrome, but realistically the problem could be anywhere along the course of the median nerve proximal to the carpal tunnel. We'll see how we analyze this situation when we come to discussing specific conditions and their EMG findings.
Median Motor Conduction
With median motor studies, there is a slightly different process involved for calculating our NCVs. The signals we obtain for motor studies are measured in millivolts rather than the microvolts for sensory studies. This means they are easier to obtain, and can be tracked over longer distances.
We start the median motor study by putting recording electrodes over the thenar eminence, then stimulate at the wrist. As we said in the previous section, at the wrist the nerve is mixed motor/sensory. By recording much larger potentials from muscle, we filter out the motor nerve conduction.
Median Nerve – Motor Conduction Study
Stimulate at wrist or elbow, record from thenar eminence