Artifacts in Neurophysiologic Intraoperative Monitoring
Aatif M. Husain, Stephanie L. Schwartz, and Emily B. Kale
Neurophysiologic intraoperative monitoring (NIOM) involves recording low-amplitude neurophysiologic signals in the operating room (OR). NIOM brings all disciplines in clinical neurophysiology together in the OR, a particularly challenging environment that can produce a lot of artifacts. Artifacts are defined as unwanted signals that arise from sources that are not the ones of primary interest (1).
Neurophysiology laboratories are carefully designed and shielded to minimize extraneous artifacts from contaminating the neurophysiologic signal that is to be recorded. In the OR, such shielding is not available; rather, the OR is an environment that is “hostile” to neurophysiologic recordings (2). There are many electrical devices, movement of personnel, other simultaneous procedures, and lack of ready access to the patient that make the neurophysiologic data acquired in the OR particularly susceptible to artifacts. Also, removing objects or disconnecting equipment from power supplies cannot be readily done in the OR, making troubleshooting and eliminating artifacts more difficult.
In this chapter, artifacts commonly encountered in NIOM will be discussed. An overview of artifacts from the perspective of NIOM will be presented first, followed by the classification of these artifacts. Thereafter, the common types of artifacts encountered in NIOM will be described and examples presented where appropriate. How artifacts are recognized, differentiated, and eliminated (or avoided) will be discussed.
NIOM involves recording many types of neurophysiologic signals. EEG, electromyography (EMG), and evoked potentials (EPs) are often recorded simultaneously. Artifacts that appear in one type of monitoring modality may be the activity of interest in another type of monitoring. For example, EMG activity in the EEG channels confounds its interpretation, but there may be simultaneous EMG monitoring to assess peripheral nerve function. Eliminating some artifacts is sometimes not possible or desirable, making managing artifacts challenging in NIOM.
Recognizing artifacts in NIOM and deciding between surgery-induced changes and artifacts is a primary responsibility of the neuromonitoring team. Artifacts can mimic significant changes and they can make recognition of these changes difficult. Misinterpreting artifacts can lead to adverse patient consequences in all of clinical neurophysiology, but it can have particularly devastating consequences in NIOM (3).
Artifacts in NIOM can be thought of as continuous and intermittent types. Continuous artifacts are ones that are continuously present with some variability in waveform shape or amplitude. Ongoing 60-Hz electrical artifact can appear as a continuous artifact while trying to record EPs. Intermittent artifacts may occur as single or runs of independent waveforms that disrupt the background activity. Bursts of EMG activity may appear intermittently while recording EEG. At times, intermittent artifacts 358may appear superimposed on continuous ones. Recognizing artifact as continuous or intermittent helps to identify the source and may help to intervene and eliminate it.
At times artifacts are called “interference” or “noise,” as they interfere with the signal of interest. In most cases, artifacts are to be avoided and eliminated when encountered. This can be done by changing the head or limb position, administering medications (like neuromuscular junction blocking drugs, paralytics), and use of filters and other amplifier settings. Some of these manipulations may not be possible during surgery. The limitations and changes in signals caused by changing amplifier settings should be appreciated. Occasionally, artifacts cannot be eliminated and should simply be recognized and mentally discounted (1).
Artifacts in NIOM, as in most of clinical neurophysiology, can be classified based on whether they are physiologic or nonphysiologic (Table 13.1). Physiologic artifacts are electrical potentials that originate from the patient from a source other than the one of primary interest. The physiological artifacts seen in NIOM are different than those seen in other types of neurophysiologic tests. In EEGs, extraocular movements and glossokinetic artifacts frequently contaminate the recording. These do not typically occur in anesthetized patients undergoing NIOM. Similarly, ECG artifact is seldom an issue in EP and EMG data, though occasionally it can be seen when recording EEG in the OR.
Nonphysiologic artifacts arise from sources external to the patient. There are many nonphysiologic sources of artifact that are unique to NIOM and not seen in other neurophysiologic procedures. There are many machines in the OR that can produce a continuous or intermittent electrical artifact. The electrocautery, patient warming devices, operating microscope, x-ray equipment, operating table, and other electrical items are only seen in the OR and produce a unique type of artifact. Sometimes a device that produces nonphysiologic artifact in one monitoring modality looks different than the artifact it produces in another modality.
There are many types of artifacts that can be seen in NIOM, and new ones are routinely discovered with the introduction of new devices and machines. The more common types of artifacts encountered in NIOM, along with examples, are discussed in the following sections.
d. ECG and pulse
e. Other physiologic signals
a. NIOM system
iii. NIOM machine
b. Other equipment
i. Electrical power lines
iii. Temperature management systems
iv. Operating microscope
v. Ultrasonic surgical aspirator
vi. Brain stimulator
vii. x-Ray equipment
viii. Surgical drill
ix. Cardiopulmonary bypass machine
x. Other sources
EMG, electromyography; NIOM, neurophysiologic intraoperative monitoring.
In NIOM, physiologic electrical signals may be artifacts or the signal of interest, depending on the modality being recorded. For example, during spinal surgery, EMG activity is considered physiologic artifact in the somatosensory evoked potential (SEP) recording, but it is the signal of interest when recording EMG activity to monitor for spinal root irritation. Thus, in NIOM it may be undesirable to completely eliminate some physiologic signals, rather it is best to manage them depending on what is being monitored. The common types of physiologic artifacts seen in NIOM are discussed in the following sections.
EMG artifact is a very common physiologic artifact encountered in NIOM. In an anesthetized patient, while there is little movement, muscle tone is still 359present, unless paralytics are used. EMG artifact is of high frequency and may be seen as a continuous or intermittent discharge (4). A continuous EMG discharge may occur when the patient’s anesthesia is inappropriately light or if there is a subconscious perception of pain. Intermittent EMG activity may be noted with unexpected movement or irritation of nerves and muscles. As noted earlier, although EMG artifact can make averaging EPs and interpreting EEG more difficult, it is the primary signal of interest when peripheral nerves or roots are at risk of injury. At times EMG may be noted with motor evoked potentials (MEPs) as well, appearing as random waveforms on the screen when an MEP is acquired. An example of this is presented in Figure 13.1.
When EMG artifact becomes excessive, elimination attempts can include moving the patient’s head or limbs if possible. Using paralytics will help, but it may affect other types of monitoring and is usually avoided. A lower high-frequency filter (HFF) can also help, but it may distort other high-frequency signals of interest (such as various EPs). When recording EPs, increasing the number of repetitions averaged may help improve the signal as well. At times, the EMG artifact must simply be recognized and discounted.
Another type of EMG artifact is seen when muscle contraction occurs with peripheral nerve stimulation that is used during SEP monitoring. Ulnar and median nerve stimulation causes contraction of hand muscles, and this can be seen as regular motor potentials on free-running EMG recorded from the same muscles. Similarly, tibial nerve stimulation causes contraction of foot muscles, and regular motor potentials are seen in the free-running EMG from foot muscles. An example of this type of artifact is presented in Figure 13.2. These discharges are easy to identify; they cannot be eliminated but should be recognized and not confused with abnormal, spontaneous EMG activity from the same muscle groups.
In many types of NIOM cases, EEG is the signal of interest. However, when recording various types of EPs, EEG activity becomes a continuous type of artifact that may reduce the reproducibility of the EPs. The frequency of EEG activity is usually between 0.5 and 30 Hz, though with craniotomies, higher frequencies may be noted. An example of EEG activity becoming artifact is when visual evoked potentials (VEPs) are recorded in OR. VEPs are highly sensitive to anesthetics, particularly inhalational ones, so total intravenous anesthesia (TIVA) with propofol is desirable. If the propofol dose is kept low, EEG activity may obscure the VEPs, as EEG is of much higher amplitude.
Eliminating all EEG activity is often not desirable or possible. Usually, reducing EEG activity involves increasing the anesthetic dose. The higher anesthetic dose may make obtaining the signal of interest more difficult. In the preceding example, EEG can be reduced by increasing the dose of propofol. However, if the depth of anesthesia increases considerably, obtaining VEPs may become impossible. In this example, the EEG artifact is best dealt with by increasing the number of repetitions averaged with each EP.
Respiration can produce a slow-frequency artifact that occurs at the same frequency as the patient’s breathing, about 5 to 15 per minute. In the OR, the patient’s respiration is controlled by a ventilator, so the respiratory rate can easily be determined. Respirations can cause mild head movements, which causes mechanically induced impedance changes in the EEG leads (5). This manifests as artifact in the EEG channels as a continuous, very slow undulating wave. When diaphragmatic or intercostal EMG is monitored with needle electrodes, respiratory artifact may appear as a high-frequency, phasic burst of activity during inspiration when the muscles of respiration contract. Although this is considered an artifact, it also is useful as it confirms appropriate placement of the recording electrodes.
Respiratory artifact often cannot be entirely eliminated, and this may not be desirable either. Movement-induced respiratory artifact noted on EEG can usually be corrected by repositioning the patient’s head if it is recognized before the patient is draped. Increasing the low-frequency filter (LFF) can also help, but it may distort other physiologic activity of interest. Respiratory artifact seen in diaphragmatic EMG cannot be easily eliminated, but it may serve a useful purpose and can easily be ignored.
Electrocardiogram and Pulse
The ECG artifact is seen as a regular, continuous artifact occurring at the same frequency as the heart rate. In the OR, it appears occasionally in EEG recordings and can be easily recognized and discounted. When ECG artifact appears in free-running EMG channels, it may be harder to recognize as typically only 1 second of EMG activity may be seen per screen (sweep of free-running EMG is usually 100 msec/division). Related to ECG artifact is pulse artifact. This occurs at the same frequency as the heart rate, though there is a slight lag in the pulse artifact compared to the QRS complex of the ECG. A pulse artifact is seen when the recording electrode is over or near a blood vessel. Though this is seen usually in EEG electrodes, it can also be seen in other modalities as well. If possible, the recording electrode should 360be moved slightly, but if that cannot be done, it is easy to ignore. An example of ECG and pulse artifact is presented in Figure 13.3.
Other Physiologic Signals
Several other types of physiologic artifacts can occur during NIOM, but they are much more common in EEG, EP, or EMG recordings. They are discussed in more detail in their respective chapters, and only a brief summary is presented here. Eye movements are a common type of physiologic artifact seen in the EEG laboratory. They produce an artifact in the frontal leads that can have a slow (slow eye movement) or fast (rapid eye movements) upslope, followed by a slow wave. Eye movements are seldom an issue when patients are under general anesthesia. Glossokinetic artifact from tongue movement produces a high-amplitude slow wave better seen in the temporal channels of the EEG. Like eye movement artifact, glossokinetic artifact is seldom a source of problems in the OR.
Sweat can cause artifacts in recording electrodes. Sweat itself can cause low-amplitude slow discharges due to its electrically charged nature. It can also cause electrode impedance issues and formation of salt bridges between electrodes, making recording signals difficult. The very slow sweat artifact can often be eliminated by increasing the LFF or by cooling off the patient. In the OR the temperature is often fairly low, and sweating is usually not a problem.
There are many possible sources of nonphysiologic artifacts that can occur during NIOM. They can be thought of as related to the monitoring system, from electrodes to the NIOM machine. Alternatively, they can be due to sources external to the NIOM system and the patient and originate from the environment, devices, and machines located in the OR.
Artifacts related to the NIOM system may come from the electrodes, wires, and electrode interface with the patient or from an NIOM machine.
Electrodes and Wires
Electrodes attached to a patient can produce many types of artifacts. Electrode artifacts are often limited to a single channel, making their identification easier. The commonest are “electrode pops.” These are intermittent, sharp waveforms. Their morphology can mimic epileptiform spikes. They can occur in runs and mimic a rhythmic discharge. These artifacts occur more commonly when electrodes of different types are combined in a single channel, such as a needle electrode referenced to a surface electrode of a different metal (4). Electrode pops and other electrode artifacts are more common when the interface between the electrode and patient is impaired, such as when the conductive paste has dried or the electrode has become loose.
Wires that connect the electrodes to the headbox can also be a source of artifacts. Movement of the wires causes electrostatic artifacts that can mimic intermittent slow waves. In the OR, personnel movement around a patient cannot be controlled, and this movement may result in wire movement-related artifact. Damage to the wire may also be a source of artifact. Such damage may introduce electrical noise from other sources. Such an example is presented in Figures 13.4A to C.
Electrode and wire artifacts are best managed before they happen. Good technique in electrode application, headbox positioning, and wrapping the wires of the electrodes can help in preventing artifacts from appearing. This is very important in NIOM as once the patient has been draped, access to the patient to adjust electrodes is very limited. Once electrode artifacts appear, checking impedance can help determine if the electrodes are at fault. Often the impedance will be high. If a redundant electrode is available, it should be substituted for the electrode not making good contact with skin. Wire movement artifact can be reduced by using a LFF. Although electrode artifact that cannot be corrected can be ignored, it often foretells problems with reliability of monitoring as the recording electrodes are likely unable to faithfully reproduce the signal of interest.
Internal noise of the NIOM machine can also produce a variety of artifacts. Such artifacts are mostly due to problems with the headboxes or amplifiers in the machine. Movement of electrical charge through the input contacts on the headbox to the various components of the NIOM machine can produce these artifacts. Modern NIOM machines are made of components of very low inherent noise, and machine-related artifacts generally do not occur unless there is a machine failure. These artifacts can be of many types and may be continuous or intermittent. Their morphology is usually not similar to that of usual neurophysiologic signals and are easy to differentiate from the activity of interest. Unlike electrode artifacts, these artifacts usually occur in many channels simultaneously. An example of artifact related to bad inputs in the headbox is presented in Figures 13.5A and B.
Unfortunately, NIOM machine artifacts are difficult to eliminate other than by replacing the offending component of the NIOM machine (headbox, 361amplifier, etc.) or using alternate parts of the same equipment (using different inputs of the headbox, if available). Because components of an NIOM machine are difficult to change during a surgery (except for the headbox), when machine-related artifacts are seen, a different machine is used if available. Regular maintenance of equipment can prevent machine failure in the OR, and biomedical checks of all NIOM machines should be conducted according to the manufacturer’s instructions.
Environmental artifact is one of the most common types of artifact seen in NIOM. Electrical line artifact is the most common, but there are many devices and equipment in the OR that can also produce environmental artifact (6).
Electrical Power Line
Alternating current (AC) in electrical power lines is a common source of artifact in NIOM. In the United States, AC is at 60 Hz, while in some European countries it is 50 Hz. Electrical power lines can produce a 60-Hz artifact in two primary ways (5). The common way this artifact appears is through electrostatic induction. Capacitance between the AC mains wiring and another conductor acting as a capacitor (such as electrodes and wires) can generate capacitive voltages. If these capacitive voltages reach the differential amplifier, they will be amplified if they are unequal in inputs 1 and 2. This will produce a 60-Hz artifact.
Electrical power line artifact can also occur through electromagnetic effects. This is caused by nearby electrical equipment, such as lamps, television screens, and other medical devices. These devices generate an electrical field that is inductively coupled with wires attached to a patient. At 60 Hz, or more frequently a harmonic of 60 Hz, such as 120 Hz, artifact is produced. An example of this type of artifact is in Figure 13.6.
The 60-Hz artifact is a very regular, continuous artifact. Its amplitude varies depending on the position of the patient in relationship to a variety of electrical equipment in the vicinity, as this will change the strength of the electrical field causing the artifact. At times, a 60-Hz or 120-Hz artifact can mimic an EMG discharge. They can be differentiated by increasing the sweep of the monitor to count the sinusoidal waveforms. Using a 60-Hz notch filter will almost entirely eliminate a 60-Hz artifact, but it will not change EMG activity or a 120-Hz (or other harmonic of 60) artifact very much. The presence of a 60-Hz artifact may also indicate poor electrode contact with the patient.
Good electrode application technique minimizes the chance of seeing excessive 60-Hz artifact. Many modern buildings have electrical lines grounded to the same earth ground, minimizing electrostatic interference. Using a 60-Hz notch filter can greatly reduce this artifact, but if this filter is applied after baseline data have been acquired, there may be a significant change in waveform latencies and amplitudes. These changes should be recognized as induced by the addition of a filter. Sometimes possible sources of electromagnetic interference can be eliminated by turning off various machines in the OR. Of course, a discussion with the surgeon and circulating nurse must occur before this is done to prevent disconnection of a vital piece of equipment. Temperature management systems and other such equipment can be turned off for at least a few minutes to help isolate the source of the artifact. If such a source is found and cannot be permanently disconnected, moving the headbox and NIOM machine as far away as possible from the offending machine may help reduce the artifact.
Electrocautery causes a large-amplitude, high-frequency discharge (7). This discharge is easy to identify due to its large size and distinct morphology. It is intermittent and occurs only when the electrocautery is used by the surgeon. To the technologist in the OR, an audible sound may be associated with this discharge. The electrocautery discharge disrupts EEG and EMG signals by saturating amplifiers, which may take a few seconds to recover and start displaying physiologic signals again. An example of this is presented in Figure 13.7. It can also overwhelm the EPs if averaging is not paused. This is a commonly encountered artifact due to the frequent, necessary use of electrocautery during surgery.
Electrocautery artifact cannot be completely avoided. Modern amplifiers that are rated for OR use have rapid recovery times so little EEG and EMG data is missed from amplifier saturation. Artifact rejection software in the NIOM machine may prevent electrocautery artifact from being averaged with the EPs. However, despite using such software, it is best to pause the averaging because if the artifact is below the threshold value of the artifact rejection parameters, it can disrupt the EP average.
Temperature Management Systems
Temperature management systems, such as the Bair Hugger, help prevent the patient from becoming hypothermic during surgery. These systems can introduce a significant artifact in NIOM. The artifact produced by the temperature management system is a continuous 100-Hz artifact (8). Elimination of the artifact when the temperature management system is turned off positively identifies it as the cause. Averaging SEPs and brainstem auditory evoked potentials (BAEPs) can be very difficult with this type of artifact as the amplitude of the artifact is often not large enough for it to 362be eliminated by the artifact rejection software (it is within bounds of the artifact rejection parameters).
Artifact from the temperature management system can be reduced by moving electrode wires away from the device. Moving the NIOM equipment away from the temperature management system is also helpful in reducing the artifact. Increasing the number of repetitions in the EP average may help as well. If absolutely necessary, the surgery and anesthesia teams may allow it to be turned off, at least during critical periods in surgery.
The operating microscope can be a major source of artifact in NIOM. Different operating microscopes produce different degrees of artifact. The artifact is usually of high frequency, typically continuous, and a harmonic of the 60-Hz electrical power line artifact (9). SEPs and BAEPs are most affected by this type of artifact. An example is shown in Figures 13.8A and B.
Elimination of artifact induced by the operating microscope may be challenging. It is not removed by the 60-Hz notch filter as it has a harmonic frequency. Ensuring that the operating microscope is grounded to the same earth ground and has no leakage current is important. Moving the NIOM machine to a position in the room away from the operating microscope may help.
Cortical stimulators, such as the “Ojemann cortical stimulator,” are used to deliver biphasic electrical stimuli, usually at 50 to 60 Hz, directly to the cortex for 1 to 5 seconds to disrupt underlying neural networks and validate eloquent cortical function. It is used after a craniotomy has been performed when mapping the brain for language and motor areas of the cortex are desirable. Simultaneous electrocorticography (ECoG) is often recorded to monitor for after discharges. When stimulation is performed at 60 Hz, a 60-Hz artifact is seen in the ECoG, and this may saturate the EEG amplifiers. Modern amplifiers recover quickly and register ECoG almost immediately after stimulation ceases. The amplitude of the artifact depends greatly on the position of the recording electrode in relation to the stimulator.
Although this artifact is due to brain stimulator activation, it is not necessary to eliminate. In fact, seeing the stimulation artifact alerts the NIOM team about when stimulation is performed so that they can be vigilant about reporting after discharges. Figure 13.9 shows an example of such an artifact.
Ultrasonic Surgical Aspirator
An ultrasonic surgical aspirator allows simultaneous tissue fragmentation, suction, and irrigation and is commonly used in neurosurgical and many other surgical procedures. It produces a very high-frequency, low-amplitude continuous artifact. Because the ultrasonic surgical aspirator artifact is of low amplitude, it is not eliminated by the artifact rejection software (9). If averaging of SEP and BAEP is not paused during the use of the aspirator, the artifact overwhelms the EP average and greatly reduces the amplitude of the signal of interest.
The ultrasonic surgical aspirator artifact may not be possible to eliminate entirely. Pausing the averaging during EPs is the most effective way of reducing this type of artifact. Repositioning the NIOM equipment and the aspirator may help as well.
Many ORs have x-ray equipment, such as the C-arm, that is intermittently used during surgery. The C-arm produces high-frequency artifact at 100 Hz, similar to the temperature management systems (8). This artifact may be present continuously while the C-arm is being used, but may also be present when it is not being used but still left on and in the OR.
Artifact from the C-arm can be reduced by moving the electrode wires away from the imaging device, and if possible, moving the NIOM equipment away as well. Increasing the number of repetitions averaged also helps. Finally, turning the C-arm off and disconnecting it from power when it is not in use will help, if allowed by the surgical team.
Surgical drills are used in many neuro and orthopedic surgeries. When in use, the surgical drill produces a high-frequency, continuous artifact. Typically, the amplitude is not very large, however it is much larger than SEP and BAEP waveforms.
Artifact from the surgical drill is very hard to eliminate. Because of its low amplitude, it is typically not eliminated by artifact rejection software. The best way to eliminate this artifact is to avoid it by not averaging EPs while drilling is ongoing. If averaging is not suspended when the drill is being used, the EP average can be overwhelmed. This is particularly the case in cerebellopontine angle surgeries. When the surgical drill is used during exposure, if averaging BAEPs is attempted, there may appear to be a reduction of amplitude compared to baseline. This can lead to false alerts to the surgical team.
A surgical drill placed directly on a limb may touch the needle electrodes in the muscle underneath. This can produce a high-frequency discharge in the EMG channels and can mimic spontaneous activity. Removing the drill will eliminate the artifact. An example of this type of artifact is presented in Figure 13.10.
363Cardiopulmonary Bypass Machine
A cardiopulmonary bypass (CPB) machine is often used in aortic surgeries. A CPB machine is often used when EEG and/or EP monitoring is being performed. It can produce many types of artifacts that can interfere with NIOM, including electrical equipment interference, mechanical movements, static electricity, and piezoelectric effects (10). When EEG is being used to determine electrocerebral inactivity (ECI) in cases of deep hypothermic circulatory arrest, the artifact can make determination of ECI challenging. The high sensitivity of the EEG recording makes the artifacts more prominent (6). Additionally, the artifact can look like ECG, making it seem as if cardioplegia has not occurred. This latter issue is more problematic for the surgical and anesthesiology teams, but the NIOM team should also be aware of this effect.
Several techniques can be used to help eliminate CPB artifact. Moving the EEG or EP machine to the opposite side of the room as the CPB machine may help reduce the artifact. Proper grounding of the NIOM equipment, the CPB machine, and the patient can help reduce the electrical artifact as well.
There are many other devices and electrical equipment in the OR that can produce nonphysiologic artifacts during NIOM. Though these may be encountered less often than those already discussed, they are no less important to identify and try to eliminate.
Devices like the operating table may produce an electrical artifact which may be challenging to identify. Consumer electronics, such as cellular phones, are now increasingly used in the OR. Sometimes they can produce an artifact that can be identified by virtue of it appearing only when the device is in use. Interventions by the surgeon in the surgical field, such as hammering hardware, can cause intermittent artifacts as well. Generally, continuous artifacts are harder to identify and eliminate with averaging than intermittent ones. Intermittent artifacts can often be isolated by virtue of their relationship to a particular event or piece of equipment being turned on, cycling, or brought into the room.
Artifacts are ubiquitous during NIOM. Avoiding artifacts is better than eliminating them; this can be done most effectively with good electrode application and setup for the monitoring. When artifacts appear, identifying them and trying to eliminate or minimize them is important to preserve the integrity of the underlying neurophysiologic signal. In some instances, artifacts cannot be eliminated and do not impede recording the activity of interest. In these cases, it is best to recognize the artifact and mentally adjust interpretation appropriately.
EMG, electromyography; MEP, motor evoked potential.