EMG Technique

Purpose

To learn how to set up electromyography (EMG) recordings for TMS and troubleshoot problems with biological and electrical noise.

Videos demonstrating how to identify the ECR and FDI muscles, how to perform skin preparation, and how to check for electrical impedance can be found here.

Overview

It is important to obtain good quality EMG traces that are free from noise issues because noise can make it harder, or impossible, to detect MEPs which means a MEP status cannot be determined.

Obtaining good quality EMG recordings involves several key steps.

Ensure the EMG parameters are correct
Set up the equipment so the EMG recording is triggered by the TMS unit.

You should be able to see the EMG trace before and after the trigger, and the trigger should occur at 0 ms.

Ideally the EMG trace begins at least 20 ms prior to the TMS stimulus and ends at least 50 ms after the TMS stimulus.

Securely position surface EMG electrodes in a standardised way
Prepare the skin over the muscles of interest with the following steps to improve the quality of the EMG traces:

Shave to remove any hair
Clean the skin with an alcohol wipe
Lightly abrade the skin with prep tape
Position the electrode so the center is on the skin you have prepared, over the belly of the muscle
Securely stick the electrode to the skin
Tape can be used to secure the electrodes once they are placed to ensure good contact with the skin

For Extensor Carpi Radialis (ECR), two electrodes can be positioned in a belly-tendon montage. Put one electrode over the muscle belly and the other over the dorsal surface of the wrist, centered between the radial and ulnar styloid processes.

You can ask the patient to extend their wrist with their fist closed if they are able. This can help identify the ECR muscle body and isolate it from the extensor digitorum muscle body. If the patient is unable to extend their wrist then the ECR muscle belly can be found by palpating the large brachioradialis muscle belly that runs along the radial (lateral) border of the forearm, and then rolling your fingers off to the extensors, which should lead to palpating the ECR muscle belly. The ECR muscle belly should be along an imaginary line running from the lateral epicondyle of the humerus to the styloid process of the radius.

For First Dorsal Interosseous (FDI), one electrode can be placed over the muscle belly and one on the dorsum of the hand. The FDI muscle body is located slightly proximal to the webbed space between the thumb and index finger on the dorsal side of the hand. It is the thicker, fleshy part of the web between the thumb and index finger, and can be palpated by having the patient abduct their index finger if possible.

If you are using a ground electrode then it should be positioned over a bony area such as the lateral epicondyle of the humerus.
If you’re using a ground electrode strap then it should be placed around the arm just proximal to the elbow. The metal pin should be visible and facing outwards and not in contact with the participant’s skin.

Electrodes should not be placed over areas of broken skin.

Electrodes need to maintain contact with the skin throughout the session. Lost contact can occur if the patient has perspiration on their hands. Electrodes can be taped in place to help maintain good contact if needed.

Securely connect the electrodes to the EMG unit
Connect the electrodes to the EMG unit via cables, ensuring they are securely attached at both ends.

Channel 1 is ECR
Channel 2 is FDI

Noise-Free EMG Traces

Below are some examples of EMG traces that are free from noise issues. This is what you should be aiming for with your own EMG traces.

Figure A: In the above image you can see the ECR (top) and FDI (bottom) EMG traces are flat with no noise present, and there is an MEP present in both traces.

Figure B: Both EMG traces in the above image are free from noise issues, which you can tell from them being completely flat. No MEPs are present.

Figure C: The above image shows numerous very small MEPs in the FDI with amplitudes of ~20 microvolts. These MEPs could only be detected because the EMG is free from noise as even a small amount of biological or non-biological noise would obscure these MEPs and could lead to an incorrect MEP status determination.

Biological Noise

You need to check the EMG signal is free of biological noise. There are two main sources of biological noise: muscle activation and sporadic motor units firing.

Muscle activation
Muscle activation means the ECR and/or FDI are active when stimulation occurs. This interferes with the quality of EMG traces and make it difficult to identify MEPs, which is why you should aim to have your patient’s hands and arms completely at rest during TMS. You will need to monitor the EMG signal in each trial throughout each session and try to eliminate any noise that arises due to muscle activation. If the patient is performing bilateral facilitation then intentional muscle activation is expected in the EMG traces and should not be troubleshooted.

The key feature to distinguish muscle activation from non-biological noise is that muscle activation noise does not have a regular pattern in the EMG trace, as you can see in the examples below.

If you notice muscle activation noise in the EMG traces then these troubleshooting suggestions to help your patient stay relaxed and at rest:

Ensure that the patient’s hand and arm are resting relaxed somewhere low like a pillow on their lap or the bed rather than on a table or similar hard elevated surface. Elevated surfaces can lead to muscle activation, particularly in the ECR muscle.
Pick the patient’s arm up off the pillow, with one hand under their wrist and another hand under their elbow. You can gently move the arm around in space, asking the patient to just relax and give you the full weight of their arm – you can feel whether they relax and let you have the weight. Then just gently drop or place the limb back on the pillow with the instruction to drop their shoulders.
Say to the patient: Just take a deep breath, let it out, and drop your shoulders.
Another possible instruction: Imagine you’re relaxing in a spa pool, and your arms are just floating in front of you supported by the warm water.
Try positioning the patient with the shoulder internally rotated so the forearm is positioned across the front of the body, rather than having the arm pointing straight out away from the body.
Trying different hand positions on the pillow is useful particularly if the FDI is activated, such as having the hand half-open with fingers flexed, rather than resting palm-down with the fingers extended.
Give the patient a short break from TMS if you have been delivering stimuli for several minutes.
Make sure the upper limb that is not being recorded from is also at rest. Muscle activity in one upper limb increases activity in the contralateral homologous muscle.
If all else fails, you can also try letting the arm hang down at the patient’s side, over the edge of the chair (much harder to do if they are in a bed). This lets the arm just hang under its own weight, but care is needed if the patient has an unstable shoulder.
If none of this works and you’re still seeing some low level muscle activation noise then you’ll need to just keep going.

Figure D: The ECR muscle EMG trace (top) displays biological noise due to the ECR being active. The noise in the ECR trace can be identified as muscle activation because it does not have a regular pattern. No MEP is present in the ECR trace. No noise problems are present in the FDI muscle EMG trace.

Figure E: The ECR EMG trace (top) displays biological noise because the ECR is activated. This noise can be identified as muscle activation because it does not have a regular pattern. An MEP is present in the ECR trace and more information about identifying MEPs in the presence of biological noise is provided in the Challenging MEP identification module.

Muscle activation noise is also present in the FDI EMG trace but no MEP is present. The noise in both traces can be identified as muscle activation because it does not have a regular pattern.

Figure F: The ECR EMG trace (top) displays biological noise which indicates the ECR is activated. The noise in the ECR trace can be identified as muscle activation because it does not have a regular pattern which electrical noise typically does, as seen in the below sections.

A low level of muscle activation is present in the FDI trace, as well as an MEP with a latency of approximately 20 ms. Sporadic motor unit firing can be seen in the fluctuations in the FDI EMG trace at approximately -12 ms and 95 ms. This is identified as a motor unit firing because of their consistent shape and the latencies do not correspond to the expected FDI MEP latency.

Figures G-H: The two figures above show severe biological noise due to muscle activation in the blue ECR traces. There is low-level biological noise due to muscle activation in the red FDI traces of Figure G, but the red FDI traces in Figure H are largely noise-free. These figures illustrate how biological noise due to muscle activation does not have a consistent shape or rhythm between different traces. No MEPs are present in any trace which you can tell because there are no consistent deflections in the expected MEP latency windows of ~15 – 25 ms for the ECR and ~20 – 30 ms for the FDI.

Sporadic motor units firing
A second form of biological noise is caused by motor units firing sporadically. A motor unit is made up of a single motor neuron axon and all of the muscle fibers it innervates. Motor units can sometimes spontaneously fire due to random firing of the motor neuron, but this activity is usually not consciously noticed.

Motor units firing can be identified due to their consistent size and shape in EMG traces. While the shape can sometimes resemble an MEP, they usually appear at latencies which do not correspond with the expected MEP latency. ECR MEPs have an expected latency of ≈ 15 – 25 ms and FDI MEPs are ≈ 20 – 30 ms. If a motor unit is firing spontaneously then you will usually see it appear more than once in the trace.

Re-positioning a patient and ensuring their arms and hands are fully supported on pillows can prevent or reduce noise from sporadic motor unit firing.

Figure I: The fluctuations seen in the ECR EMG trace (top) at ~20 ms and ~100 ms are not MEPs. This is a motor unit firing sporadically which you can tell because of the consistent size and shape of the fluctuations, and the firing at ~100 ms which does not correspond with the expected ECR MEP latency. Re-positioning the patient and ensuring their arms and hands are fully supported on pillows may prevent or reduce sporadic motor unit firing. A similar situation can be seen in Figure V below.

No motor units are firing in the FDI trace but a small MEP can be seen at approximately 33 ms.

Figure J: The above figure displays a series of ECR EMG traces, each of which is the result of a single stimulation. An ECR MEP is present in the first two and last two traces, and these can be seen in the green circles. Throughout the traces there are a high number of motor units firing, some of which are highlighted in red circles. Although these responses vary in size, their consistent shape and appearance at latencies that don’t correspond to the expected ECR MEP latencies indicates they are motor units firing sporadically. There are many other motor units firing in these traces that haven’t been circled. The two responses in yellow are within the expected window for ECR MEP latencies, however the fact their shape is so similar to the motor units firing, and so dissimilar to the MEPs present in the first and last two traces, indicates they are likely also motor units firing sporadically.

Non-Biological Noise

You also need to ensure that the EMG traces are free from non-biological sources of noise. The two main sources of non-biological noise are electrical noise and prolonged stimulus artefacts.

Electrical noise
Noise can be generated from electrical sources, such as poor contact between the recording electrodes and patient’s skin, or electrical noise in the room from fluorescent lighting.

Electrical noise issues are characterized by having a regular pattern in the EMG trace, as you can see in the examples below.

It’s also important to recognize that electrical noise issues come in many different shapes and sizes. For example, both rhythmic wave-like patterns and abrupt, sharp deflections in the EMG traces are indicative of electrical noise issues. Although Figures K-R below all show electrical noise issues, almost none of them show the same electrical noise.

Troubleshooting suggestions for electrical noise include but are not limited to:

checking skin preparation for electrodes was adequate which includes shaving hair, cleaning with alcohol wipes, and lightly abrading using prep tape
checking the electrodes have not lost contact with the skin. Tape can be used to secure electrodes
checking electrode cables are firmly attached at the patient and the EMG unit
checking the grounding strap/electrode has not lost contact with the skin
wiping the ground strap/electrode with an alcohol wipe if it hasn’t been cleaned before use
checking whether anyone is touching the patient or the TMS trolley during the test
unplugging the bed from its electrical power supply
turning off fluorescent lighting where possible
positioning the patient further away from the TMS machine
If electrical noise appears during bilateral facilitation it is likely the patient is causing electrical noise issues by pulling on the electrodes and/or cables. Changing the patient’s posture while facilitating to prevent this happening should eliminate the noise

Electrical noise present in the traces for both muscles indicates an issue with the grounding strap/electrode or the environment. In this situation follow the suggested troubleshooting steps above.

Figure K: Both channels have a regular wave-like pattern that indicates electrical noise is present. As electrical noise is present in both channels there may be issues with EMG set-up for the grounding electrode/strap, or an environmental issue such as interference from a person touching the TMS trolley or patient, fluorescent lighting, or electrical power supply to the bed.

Figure L: Each channel is affected by a different type of electrical noise. The ECR trace (top) has a drifting baseline. The FDI trace has electrical noise with a sine-wave pattern. As electrical noise is present in both channels there may be issues with EMG set-up for the grounding electrode/strap, or an environmental issue such as interference from a person touching the TMS trolley or patient, fluorescent lighting, or electrical power supply to the bed.

Figure M: The above figure shows a low level of electrical noise present in both the blue ECR traces and the red FDI traces which you can tell by the wave-like patterns. This figure illustrates how electrical noise is often very consistent looking between traces, in contrast to biological noise due to muscle activation which does have a consistent shape as seen in Figures G-H above. Moreover, even low level electrical noise such as in figure M could obscure very small MEPs such as those present in Figure C above and lead to an incorrect MEP status determination.

Figure N: Another demonstration of electrical noise looking consistent between traces in both muscles. Electrical noise of this amplitude would definitely obscure small MEPs such as those in Figure C above.

Figure O: The above image shows consistent electrical noise issues in the blue ECR traces. However, the above image also shows that occasionally electrical noise won’t be consistent between traces such as in the green circles. These noise issues can still be identified as electrical because of the very abrupt and non-natural deflections in the EMG traces, and hopefully you can see how different these noise issues look compared to biological noise due to muscle activation in Figures D-H.

Electrical noise present in the EMG trace of one muscle suggests the noise issues are specific to the EMG set up for that muscle. Potential causes include a loose cable, poor electrode contact with the skin, or a faulty electrode. In this scenario swapping the electrode cables between muscles is the quickest way to understand the problem:

If it is a problem with the cables then the noise issue will change channels to the other muscle. In this situation using a different set of cables should resolve the issue.
If it is a problem with the electrodes then the noise issue will stay with the original muscle experiencing noise. In this case you will need to remove the electrodes, re-prep the skin areas, and attach new electrodes.

Figure P: The ECR EMG trace (top) is contaminated by electrical noise, which you can tell based on its regular pattern. As this electrical noise is only present in one channel it suggests that the issue is specific to the EMG set up for the ECR such as a loose cable, or poor electrode conduction due to lack of contact with skin, or a faulty electrode. No noise issues are present in the FDI EMG trace.

Figure Q: The FDI EMG trace (bottom) is contaminated by electrical noise which you can tell by the regular pattern. As this electrical noise is only present in one channel it suggests that the issue is specific to the EMG set up for that muscle such as a loose cable, or poor electrode conduction due to lack of contact with skin, or a faulty electrode.

The ECR muscle EMG trace (top) does not contain electrical noise but does contain biological noise due to muscle activation. You can tell this is muscle activation noise because it does not have a regular pattern.

Figure R: The blue ECR traces are perfect and free from noise, however there is very significant electrical noise in the red FDI traces. You can tell this is an electrical noise issue because of the consistent shape both within and between FDI traces. This noise is likely to be caused by a loose cable, poor electrode conduction due to lack of contact with skin, or a faulty electrode. This FDI electrical noise is so severe you’re unlikely to be able to detect any FDI MEPs that may be present.

Prolonged stimulus artefact
The second form of electrical noise is caused by a prolonged stimulus artefact. MEP identification can be difficult if the stimulus artefact of the TMS trigger has a prolonged tail that overlaps with the expected onset of the MEP. The expected MEP onset is from 15 ms after the stimulus for ECR, and from 20 ms after the stimulus for FDI. The stimulus artefact may be reduced by following the troubleshooting advice for electrical noise issues.

Figure S: There are several points of interest in the above image:

An elongated stimulus artefact is present in the ECR trace because the ECR trace does not return to baseline until ~20 ms after stimulation. In contrast, the FDI ECR returns to baseline a few milliseconds after stimulation as expected. The stimulus artefact may be reduced by following advice on troubleshooting EMG noise issues.
The fluctuation seen in the ECR EMG trace at around 50 ms is not a MEP. This is most likely a motor unit firing sporadically. It is identified by its appearance at a latency which does not correspond with the expected latency between 15 and 25 ms for ECR MEPs. Re-positioning a patient and ensuring their arms and hands are fully supported on pillows may prevent or reduce sporadic motor unit firing.
The fluctuation seen in the FDI EMG trace at around 40 ms is also not a MEP, and is also most likely a motor unit firing sporadically.

Figure T: The two red FDI traces clearly demonstrate prolonged stimulus artefacts as they only return back to baseline at ~30 ms.

The top blue ECR trace does not have a prolonged stimulus artefact, however an electrical noise issue is present due to the consistent nature of the noise. Additionally, the abrupt upwards deflection of the trace starting at ~85 ms is indicative of an electrical noise issue similar to the abrupt and oddly-shaped deflections in Figure O above.

The bottom blue ECR trace has the same electrical noise issue as the top trace, however it also contains a prolonged stimulus artefact that only returns to baseline at ~25 ms.

Figure U: A prolonged stimulus artefact is present in the FDI EMG trace because it only returns to baseline at approximately 30 ms, which is within the expected onset latency for FDI MEPs.

There is a large deflection in the ECR EMG trace at the point of stimulation 0 ms, however this is not concerning as the EMG trace returns to baseline at approximately 5 ms and is thus unlikely to affect MEP status determination.

Figure V: The stimulus artefact in the ECR is very large in the y-axis as it reaches an amplitude of ~325 microvolts, however it returns back to baseline by ~8ms so won’t interfere with MEP status determination. This would not be considered a prolonged stimulus artefact.

The yellow FDI trace contains 2 motor units. Although the deflection at ~18 ms is within the expected latency for FDI MEPs, the fact it has an identical shape and amplitude to the deflection at ~48 ms indicates they’re both likely motor units firing. Seeing deflections of similar amplitude and shape at sporadic latencies in other FDI traces would further confirm these are motor units firing. Figure I shows a similar situation of two motor units firing despite one of them being in the appropriate latency range for FDI MEPs.

FAQs

What do I do if I can’t get my patient to relax their arms and there is low-level muscle activity level in the EMG trace?
First try adjusting the patient’s position to ensure their arms and hands are fully supported on pillows. Reminding them to drop their shoulders and relax as much as possible may help. You could reinforce your instruction with a descriptive statement such as ‘let your arms and hands feel heavy and fully supported by the soft pillow beneath them’ or similar. You could also lift and support their hand and arm off the pillow, gently move it until you feel them relaxing and giving you the full weight of their limb, and then let it drop naturally onto the pillow. If these techniques don’t help you could try drawing the patient’s attention away from testing procedures, to reduce anxiety and fidgeting associated with testing. It may also be worth trying to get started with some low intensity stimulation to get the patient used to TMS, and see if voluntary activation reduces over a few minutes.

What do I do if I notice noise in the EMG trace?
Start by seeing if the issue is in one muscle trace or both.

Background noise present in one muscle trace indicates the problem is specific to the EMG set up for that muscle. If you have swapped the EMG cables between the muscles and the noise ‘moved’ to the other muscle, then checking the cable connections or trying a different set of cables can help. If you have swapped the EMG cables between the muscles and the noise ‘stayed’ with the same muscle, then it is good to check that the electrodes for that muscle:

have not lost contact with the skin
are firmly attached at both the electrode and the EMG unit ends
if noise persists then re-apply new electrodes after re-doing skin preparation

Non-biological noise that is present in the traces for both muscles indicates an issue with the grounding electrode (or strap) or the environment. Troubleshooting can include but is not limited to:

Checking the ground electrode:

has not lost contact with the skin
that cables are firmly attached at both the electrode and the EMG unit ends
if noise persists then re-do the skin preparation for the ground electrode and re-apply it

Checking the environment:

is anyone is touching the patient or the TMS trolley?
is the bed connected to an electrical power supply and can this be disconnected?
can the lighting be adjusted by turning off fluorescent lighting?

How can I tell if it is a motor unit firing or a real MEP?
Review the latency and shape of the response. If the response is well outside the expected latency window it is unlikely to be a MEP. If the response is within the expected latency window you need to check to see whether the response is also present at other times in the trace. If a motor unit is firing spontaneously, you will usually see it appear more than once in the trace. Motor unit responses will have a very consistent size and shape, and usually appear at least twice in the trace. These two features differentiate them from a MEP, which will only occur once and within the appropriate time window. Of course, you might see a MEP at the right latency in a trace that also contains motor unit firing. If a MEP response of an appropriate latency is identified, and there is motor unit firing but it is outside the latency window, MEP+ classification can be made.

Should I still attempt to complete the test if I am unable to fully resolve all noise issues?
There may be some circumstances when it is still possible to determine MEP status despite noise. For example, when biological noise is present in the EMG trace but the size of the evoked MEPs are larger than the noise. In this situation it is important that MEPs are consistently seen in most traces. However before classifying a patient as MEP-, all efforts must have been made to elicit a MEP if possible. This means that EMG needs to be noise-free, particularly within the expected latency window. If noise can’t be reduced to a minimum at the time of testing then seeking the opinion of another TMS operator to troubleshoot may be useful, or it may be necessary to reschedule another TMS session. It may be worthwhile to perform this in a different clinical environment if non-biological noise was thought to contribute to noise issues.

Quiz

Click here for a practice quiz to assess your learning on the EMG Technique used for PREP2.

Once the practice quiz has been completed with at least 70% correct you will be emailed a link to the final quiz for this section.

If you pass the final quiz with at least 80% correct you will be emailed a certificate of completion for the EMG Technique section. There is no limit to the number of attempts for the practice or final quiz.

<Back