Author: White, Diane M; Van Cott, Anne C
Date published: March 1, 2010
In electroencephalography, an artifact is defined as any recorded electrical potential which does not originate in the brain. There are two basic types of artifact, physiological and nonphysiological. Physiological artifacts are generated from electrical activity related to normal functioning of the patient's body. Common physiological artifacts include those generated by muscle, body movement, the heart, the tongue, the eyes, and sweat glands. Nonphysiological artifacts are generated from electromagnetic fields outside the body. Common nonphysiological artifacts include: 60 Hertz (Hz) artifact, electrode artifact, artifact generated from medical instruments (for example, ventilators and hemoperfusion devices), and movement in the environment.
EEG monitoring in the intensive care unit (ICU) is a valuable means of monitoring brain function in critically ill patients who may be comatose, sedated, or paralyzed. The ICU presents the perfect setting for the occurrence of artifact because of its broad array of electrical equipment and hectic medical environment. In the ICU, technical EEG problems arise that are beyond those experienced in the standard EEG laboratory. There is potential for physiological artifacts as well as nonphysiological artifacts, at times making EEG interpretation difficult, if not impossible. What may seem initially to be cerebral activity, with further investigation proves to be artifact. Conversely, what may seem at first to be an obvious artifact may prove to be cerebral activity. It is essential for a technologist to: a) recognize artifact appearing in the EEG, b) have the knowledge of the many different causes of artifact, and c) know what steps to take to prove, beyond a doubt, that the EEG abnormality is an artifact. The EEG technologist is the right hand of the electroencephalographer and must be able to answer the question "Is it cerebral or is it artifact?" before it is asked. The EEG recording should never end before that question is answered.
It is the goal of this review to help the EEG technologist recognize artifacts that are commonly encountered in the ICU. This paper will review the causes of artifacts, their appearance on the EEG, and how to logically and systematically evaluate a pattern that may be an artifact.
EMG and Involuntary Body Movement Artifacts
Electromyographic (EMG) activity is electrical activity generated in muscles. EMG artifact on EEG most frequently occurs from the contraction of various muscles in the scalp, face, jaw, and neck and is most often seen when the patient is awake. EMG artifact can obscure the recording of electrical brain activity and can be seen intermittently or continuously throughout the test in one or more channels on the EEG. EMG artifact often occurs when the patient is tense or anxious and has difficulty relaxing or holding still. In the ICU this type of artifact is commonly caused by the patient biting down on the endotracheal tube which connects the patient to the ventilator. EMG artifact can simulate fast mid- voltage irregular single or serial spikes (Figure 1). This activity can simulate cerebral beta activity with the use of the high frequency filter, especially when used in combination with the 60 Hz notch filter.
Attempts to eliminate muscle artifact are often futile, especially when the patient is not able to follow commands. The technologist should make sure the patient is not cold and that his/her head is resting comfortably. Try to interact as little as possible with the patient. In some cases relaxation will come eventually, so be patient and record as long as needed to get an interpretable recording. Occasionally sleep activity might be all you can successfully record that is free of muscle artifact.
Involuntary body movement artifacts in the ICU are often related to body twitching, tremor, tics, and tonic-clonic activity. Body movements produced by the face, mouth/tongue, extremities, or trunk can cause an electrode or electrode wires to move. Such movements might occur during focal or generalized seizures or may be independent of seizure activity. Involuntary body movement artifact can be recorded in one or more channels during EEG monitoring. Rhythmic movements, like a tremor not associated with seizure activity, can cause rhythmic artifact on the EEG recording that can resemble focal seizure activity. Monitoring involuntary movements of the face, arm, or leg with two electrodes referred to each other can be correlated simultaneously with the artifact on the EEG recording.
If the patient is intubated and unresponsive, drug induced muscular paralysis might be an option. Any muscle artifact, EMG or movement, can be eliminated with neuromuscular-blocking drugs (for example vecuronium bromide, trade name Norcuron«) making it possible to evaluate brain activity more accurately (Figures 2 and 3). Muscle paralyzing drugs have no known effect on consciousness and therefore do not alter brain function like sedatives. Since muscles of respiration are affected, only patients on mechanical ventilation are candidates for such an approach. Maximum neuromuscular block occurs in about three to five minutes but gradual disappearance of muscle and movement artifact on the EEG may be noted approximately 30 seconds following injection with a neuromuscular-blocking drug. Some ICU providers will administer a anxiolytic agent along with a neuromuscular agent to prevent potential patient distress related to generalized muscle paralysis.
With the advent of portable video cameras, digital video monitoring in the ICU is a practical and efficient way of capturing clinical symptoms prior to paralysis that may be later reviewed by the electroencephalographer. To attenuate muscle activity instrumentally, when all else fails, lowering the high frequency filter may prove useful.
Electrocardiogram (EKG) artifact is a common physiological artifact produced by the electrical activity of the heart. The presence of EKG artifact on an EEG tracing is related to the recording of the electrical field of the heart potential over the surface of the scalp. EKG artifact typically has a spike or sharp morphology and can be confused with epileptiform activity to the untrained eye.
The heart can be viewed as a dipole at any instant during cardiac depolarization because some myocardial cells in certain regions are negatively charged, while those in the other regions are positively charged. These polarities reflect the fact that the outside of a depolarized cell is negatively charged, whereas the outside of a resting cell is positively charged. Referential ear electrodes, Al and A2, are often in the field of the cardiac dipole. On a referential montage with A 1 or A2 in input 2 of each channel, EKG artifact deflections are initially upgoing in channels in which electrodes are referenced to Al which is positively charged from the cardiac depolarization, while downgoing in channels referenced to A2 which is negatively charged from the cardiac depolarization (Figure 4).
There is no way to completely eliminate EKG artifact during the EEG recording. In some instances turning the patient's head will change the EKG field over the head and lower the amplitude of the artifact. Routinely recording EKG simultaneously during the EEG is the best solution for this and other cardiac -related physiological artifacts. EKG rhythm can be recorded by placing two EEG electrodes on any noncephalic part of the body. Refer the EKG leads to each other (input 1 to input 2) to monitor the EKG activity. The further apart the electrodes are placed on the body the greater the EKG amplitude. This is a simple and accurate way of differentiating EKG artifact from sharper activity of cerebral origin. In addition, the use of a bipolar montage rather than a referential (A1/A2) montage usually reduces the voltage of the EKG artifact.
A pacemaker is a small device, about the size of a wristwatch, that helps to regulate heart rate. A pacemaker is made up of leads (thin flexible wires) and a generator (battery). The pacemaker is implanted under the skin just below the collarbone and can be easily visualized. Cardiac pacemaker artifact is caused by the electrical stimulus delivered to the heart by the internal pacemaker. This artifact can be seen on a regular or intermittent basis depending upon the type of cardiac pacemaker implanted in the patient. Cardiac pacemaker artifact has a repetitive spike-like morphology and may be misinterpreted as a polyphasic interictal epileptiform discharge.
When applying leads to record EKG make a conscious effort not to place the electrode on the pacemaker itself since the pacemaker battery generates a high frequency, large amplitude pulse and will contaminate the EEG recording. Placing the electrode away from the pacemaker will not eliminate the artifact but rather keep artifact to a minimum.
A cardioballistogram artifact appears as widespread rhythmic delta activity on scalp EEG and may be misinterpreted as generalized slowing of cerebral activity. Again, a channel devoted to EKG is best to monitor these artifactual slow waves that are time locked to cardiac depolarization. The waves occur at the same frequency as the EKG, but are slightly delayed, beginning just after the QRS (the highest amplitude and sharpest component of the EKG rhythm) complex. Cardioballistogram artifact on EEG is typically an issue when there is marked suppression of background EEG activity or electrocerebral silence (ECS).
A cardiac arrhythmia artifact is a physiological artifact caused by any variation from the normal rhythm of the heart beat. EKG artifact associated with cardiac arrhythmia may be misinterpreted as epileptiform abnormalities or focal slowing (Figure 5). Every precaution must be taken to avoid such errors of interpretation. As with the other physiological artifacts associated with cardiac activity, the simplest and most reliable way to avoid misinterpretation is to record EKG simultaneously during the EEG.
Pulse artifact can occur when an electrode is placed over a pulsating artery or tissue. The pulse artifact can simulate focal slowing of cerebral origin. Both the frequency and form of the waves can help identify this artifact. Like the other cardiac related artifacts, a direct relationship exists between the EKG and the pulse waves. The rhythmic slowing caused by the pulse wave does not occur simultaneously with the QRS complex of the EKG, but is typically delayed by about 200 to 300 msec after the QRS complex. A pulse artifact is easily identified by touching the electrode producing it. In this way, the technologist can confirm the movement of the electrode while simultaneously altering the appearance of the artifact. In cases when the pulse artifact originates from only one electrode, the artifact can be eliminated or reduced by moving the electrode slightly away from the pulsating tissue. Repositioning the patient's head may also prove to be helpful.
Glossokinetic Potential Artifact
Glossokinetic potential (GKP) artifact occurs with tongue movement because the tip of the tongue is electrically charged and is relatively more negative than the base of the tongue. Movement of the tongue dipole will change the electrical field around the mouth and jaw. At times the field of GKP artifact can spread to the upper part of the face. The artifact produced has a broad potential field that drops in amplitude from frontal to occipital head regions. The amplitude of the potential is greatest near the tongue and is detected best in the leads located near the mouth including lip leads, infraorbital leads, and Fpl/Fp2.
GKP artifact appears as burst of generalized slow wave activity on scalp EEG which may be misinterpreted as being cerebral in nature. With a cooperative patient, this physiological artifact can be reliably reproduced by asking the patient to repeat words that cause significant movement of the tongue such as "lalala", "Tom Thumb", or "lilt" and stopped by asking the patient not to talk or move the tongue (Figure 6). Chewing, sucking, sobbing, and hiccups, frequently seen in infants and young children, can cause GKP artifact that resembles spike-wave or other transient discharges.
In an uncooperative patient or in the ICU setting, it is often impossible to eliminate tongue movement and monitoring for GKP is imperative. An EEG channel devoted to monitoring tongue movement with electrodes placed above and below the patient's lips will correlate with the artifact. Simultaneously, the technologist should note whenever the clinical symptoms occur by indicating on the EEG that the patient is moving his tongue.
An electroretinogram (ERG) artifact is recorded during photic stimulation. It is the result of the response of the eyes' retinal cells to the light. ERG artifact is maximal in frontopolar electrodes (FpI and Fp2), since they are closest to the eyeballs. ERG is most evident during ECS recordings since cerebral activity is suppressed and high recording sensitivities are used, but may be recorded in normal individuals at standard sensitivities. An ERG artifact can be misconstrued as cerebral activity and potentially epileptiform discharges. ERG can be noted at all or some flash frequencies, is usually bilateral, and is time locked to the photic stimulus (Figure 7).
This physiological artifact will disappear when the light source to the eye is blocked by covering the patient's eye during photic stimulation (Figure 8). Both left and right eyes should be tested separately by covering the eyeball with an opaque material and simultaneously documenting the maneuver on the EEG tracing. Be careful to only cover the eye; not the FpI or Fp2 electrodes. If the artifact does not disappear when the eye is covered, consider covering FpI and Fp2 electrodes to detect a photo cell artifact (see later section "Electrode Artifacts").
In the ICU, sweat artifact is most often recorded in patients with high body temperature associated with fever. Perspiration causes slow shifts of the electrical baseline by changing the impedance between the electrode and the skin. In addition the active sweat gland itself produces slowly changing electrical potentials that are recorded by the EEG electrodes. Sweat artifact almost always appears in more than one channel, but can be lateralized or asymmetric. Sweating causes long duration slow wave artifact that is smooth in outline and of high amplitude particularly in the temporal and frontal electrodes (Figure 9), and may be misinterpreted as slowing associated with cerebral dysfunction.
Reduction of sweat artifact can be produced by cooling the patient. This can be achieved by drying the scalp with a clean dry cloth, reapplying the electrodes, and placing a cold damp cloth over the patient's forehead and front of neck or wiping the patient's face. It may be necessary to repeat this process as often as necessary to keep artifact to a minimum. Keep the room as cool as possible or, if available, use a cooling blanket on the patient.
Alternating Current (60 Hz Artifact)
The most common nonphysiological artifact is 60 Hz artifact. 60 Hz interference is caused by an alternating current which supplies power to electrical wall outlets. The waveform of an alternating current power circuit is a sine wave. The number of times an alternating current repeats a full cycle in one second is the frequency of the current and the maximum voltage of the current is its amplitude. 60 Hz artifact reverses directions cyclically - changing from positive to negative and back again - 60 times a second.
In the ICU, 60 Hz artifact comes from the electrical equipment in the patient's room (Figure 10) including such devices as: the electrically powered bed, the mechanical ventilator, intravenous infusion devices, Sequential Compression Devices (SCD) or Flowtron« Excel machines, EKG monitors, dialysis machines, fluorescent lights, and heating/cooling lamps or blankets. 60 Hz artifact can sometimes be traced back to the EEG jackbox cable if it is in contact with the floor or any other power cables (including the power cable to the EEG instrument).
Poor electrode contact associated with inadequate skin preparation, defective cables, lead wires, or faulty grounding can also cause 60 Hz artifact. There is a strong likelihood of electrode instability and high impedances with prolonged recordings in the ICU. Because of high or mismatched impedances there is a failure of common mode rejection by the amplifiers. In this situation, exogenous artifacts (i.e., 60 Hz) can be amplified and detected on the screen in one or more channels. 60 Hz artifact produces a "fuzzy" appearing baseline and can sometimes be confused with muscle artifact or fast cerebral activity. This artifact can be better identified by increasing the display (paper) speed to 60 mm/sec (Figure 11).
If electrode impedances are low (5K ohms or less) and 60 Hz artifact is still present, begin to unplug individual electrical equipment from the outlet, one plug at a time, and make a note of it on the EEG. Continue to unplug the equipment until the source of the 60 Hz is identified and the artifact is eliminated. It is always best to get permission from someone on the ICU staff before unplugging equipment associated with patient care. Never unplug the mechanical ventilator under any circumstance.
60 Hz can sometimes be ehminated simply by not letting the jackbox cable touch the floor. Use only as much cable as necessary to reach from the EEG instrument to the jackbox. Occasionally 60 Hz can easily be eliminated by keeping cables from touching each other. If all else fails and 60 Hz artifact persists even after making every effort to identify and remove the source of the 60 Hz, the 60 Hz notch filter can be used as a last option. The notch filter attenuates activity at 60 Hz and attenuates adjacent frequencies less extensively. It is important to remember that when the notch filter is used other faster frequencies, including cerebral activity, within the range of the filter is also attenuated. Hence, an epileptiform spike may have its displayed amplitude lowered because some components of the spike have frequency characteristics at or near 60 Hz.
Electrode Artifacts (From Surface Electrodes)
EEG surface electrodes are metal discs secured directly to the scalp with a conductive paste. A wire is attached to each electrode connecting the electrodes to the inputs of the EEG amplifiers. Any interruption in this path, from the scalp to the jackbox, can produce an EEG electrode artifact. Electrode artifacts can have different appearances depending on the site and type of the disruption, but are always confined to the channels that have the electrode in either input 1 or input 2.
Poor electrode contact can cause instability of impedance which leads to sharp or slow waves of varying morphology and amplitude that have the potential to be confused with focal cerebral slowing or even rhythmic activity resembling the beginning of a focal seizure. When there is an abrupt change in the electrode impedance, a sudden potential appears causing an electrode "pop." An electrode "pop" has a characteristic morphology of a very steep rise and shallow fall resembling a direct current (DC) calibration signal (Figure 12). An electrode "pop" can simulate a focal spike or sharp wave on the EEG.
Electrode artifact associated with lead wire movement has a more disorganized morphology and can be confused with generalized slowing if all electrodes are affected or focal slowing if only certain electrodes are involved. Lead movement electrode artifact occurs at the frequency of the motion of the wires.
Electrode artifact is primarily caused by improper contact of the electrode on the skin, a broken electrode wire, lead movement, or a loose electrode wire within the jackbox. Electrode artifact is likely to occur if the skin is not well prepped, if not enough conductive paste is used, or if the paste dissipates. A break in the electrode wire can occur from normal daily usage or during the recording if it is pulled or gets caught. The electrode wire can be loose in the jackbox if it is not plugged in all the way or if the wire becomes displaced.
The job of the technologist is to learn to differentiate between cerebral activity and electrode artifact. If there is questionable activity on the EEG that seems to be seen in one channel or channels with one common electrode, corrective measures must be taken. Reprep the area and reapply the same electrode with electrolyte paste. If reapplying the electrode doesn't correct the artifact, replace with another electrode you know is good or try a new one. Finally, make sure the electrode is properly plugged into the jackbox.
Photo Cell Artifact
During photic stimulation high impedance in electrodes over the frontal regions may cause an unusual artifact. Each flash causes a minute photochemical reaction which, in the presence of high impedance can cause the electrode to act as a photocell. Thus in the involved channels, a brief spike-like transient appears simultaneously with the flash.
A photo cell artifact is a time locked change in the potential of EEG in a single electrode during photic stimulation. It is easily identified as it disappears if the light is blocked from the electrode in question. When proving this is a photocell artifact, be careful to cover only the electrode not the eyeball.
Mechanical Ventilator Artifact
Mechanical ventilation is a method to mechanically assist or replace spontaneous breathing. Mechanical ventilation artifact is caused by switching magnetic fields within the ventilator motor and by the movement of the electrodes or leads as the body is moved by the device. The artifact occurs with the motor's activity, thus it may be constant or intermittent depending on how the ventilator is programmed.
Mechanical devices such as ventilators usually produce artifacts with slower components than other ICU electrical devices. The mechanical ventilator can cause bursts of rhythmic high amplitude slow waves maximal over the frontal leads bilaterally. The artifact can simulate a burst suppression pattern of cerebral origin, especially if the background EEG activity is suppressed. Ventilator artifact on the EEG can be directly correlated with mechanical breaths per minute ventilator settings or with the patient's observed respirations.
Occasionally, just repositioning the patient's head will eliminate or reduce the artifact. If the artifact cannot be eliminated it should be monitored by placing one electrode above and one electrode below the lips or by carefully attaching two electrodes to the ventilator tubing or dial. These electrodes can be referred to each other in a separate channel on the EEG. As a last option, the ventilator can briefly be disconnected from the endotracheal tube by the respiratory staff to determine if the activity is related to the ventilator.
Movement Artifact in the Environment
Movement of persons in the vicinity of the patient generates artifacts usually of a capacitive (the ability of a body to hold and electrical charge) or electrostatic origin (a sudden and momentary electric current that flows between two objects at different electrical potentials caused by direct contact or an electrostatic field). It is likely to occur as a result of movement of the patient by the nurse or doctor or movement of other personnel or family members near the patient's bed including the EEG technologist or the electroencephalographer. The morphology of the artifact on the EEG varies with the motion or activity of the person in close proximity of the patient.
It is important for the technologist to be able to recognize the correlation between what is going on in the room and what is being recorded on the EEG. It is helpful if the technologist asks the person in the room to repeat the activity or motion to reproduce the artifact on the EEG and simultaneously notes it on the recording. To eliminate this type of artifact, it is best to prevent people from moving in the room, often a daunting task in the ICU.
Prisma« Hemoperfusion Device Artifact
The Prisma« continuous renal replacement system is a type of hemodialysis that removes, filters, and returns the blood slowly and continuously in patients with kidney failure. The Prisma« device has rotary pumps which produce a unique EEG artifact during dialysis. The Prisma« device induces generalized saw-toothed waveforms in the range of 5.5 to 11 Hz that correlate with the movement of the rotor within the device. The unique artifact is characteristic, but could be easily misinterpreted as an abnormal cerebral rhythm if one is not aware of it. Prisma« EEG artifact resembles rhythmic temporal theta burst of drowsiness (RTTD), a benign EEG variant (previously called psychomotor variant) in frequency and morphology. Unlike RTTD, Prisma« artifact is located more posteriorly. It can also be misinterpreted as an epileptiform discharge. The artifact should be suspected when invariant sawtoothed rhythms appear beyond the temporal and occipital lobes and contaminate the electrocardiographic channel.
When the Prism« device is paused (rotors stopped), but not switched off, the rhythmic waveforms disappear. When the Prism« is reactivated the waveforms reappear. This artifact can be verified by its disappearance on stopping and reemergence with starting of the rotary pump action of the device (of course, this should be done in conjunction with dialysis staff).
EEG monitoring is a valuable way of acquiring additional information about brain function in critically ill patients. However, in the hostile environment of the ICU obtaining quality records can be challenging. This paper was written to educate the new or inexperienced technologist. All questionable EEG patterns or activity must be investigated. It is the responsibility of the technologist to know how to identify and differentiate artifact from cerebral activity before disconnecting the electrodes from the patient. The electroencephalographer should also be prepared to assist the technologist in finding a solution for more difficult challenges at the time of the recording. Every effort should be made for an accurate interpretation to avoid further unnecessary tests, administration of inappropriate drugs, or prolonged hospitalization.
Brenner RP. EEG on DVD - Adult: An Interactive Reading Session. New York: Demos Medical Publishing; 2007.
Brittenham DM. Artifacts. In: Daly DD, Pedley TA (Editors). Current Practice of Clinical Electroencephalography: Second Edition. New York: Raven Press; 1990, p. 88.
Fisch BJ. Fisch and Spehlmann's EEG Primer: Basic Principles of Digital and Analog EEG: Third Edition. New York: Elsevier; 1999, p. 107-21.
Katz AM. Physiology of the Heart: 4th Edition. Philadelphia: Lippincott Williams & Wilkins; 2005, p. 441.
Klass DW. The continuing challenge of artifacts in the EEG. Am J EEG Technol 1995; 35(4):23969.
Saunders MG. Artifacts: activity of noncerebral origin in the EEG. In: Klass DW, Daly DD (Editors). Current Practice of Clinical Electroencephalography. New York: Raven Press; 1979, p. 37-67.
Stern JM, Engel J Jr (Editors). An Atlas of EEG patterns. Philadelphia: Lippincott Williams & Wilkins; 2004, p. 61.
Tyner FS, Knott JR, Mayer WB Jr. Fundamentals of EEG Technology: Basic Concepts and Methods; Volume 1. New York: Raven Press; 1983, p. 281-96.
UK HealthCare. Electrophysiology: pacemakers, ICDs, and ablation fact sheet, 2007. On the Internet at: http://www.ukhealthcare.uky.edu/publications/healthfocus/fact_sheets/ electrophysiologyfst.htm Accessed September 2009.
Young B, Osvath L, Jones D, Socha EA. A novel EEG artifact in the intensive care unit. J Clin Neurophysiol 2002; 19(5):484-86.
Young GB, Campbell VC. EEG monitoring in the intensive care unit; pitfalls and caveats. J Clin Neurophysiol 1999; 16(1):40-45.
Diane M. White, R. EEG T.1 and Anne C. Van Cott, M.D.1,2
1 Veterans Administration Pittsburgh Healthcare System
2 University of Pittsburgh