A major problem in critical care EEG performance and interpretation is the proper identification and elimination, when possible, of artifacts. EEGs performed in the ICU are often contaminated by artifacts arising from monitoring equipment, life support systems and personnel. From the viewpoint of an electroencephalographer or technologist, an artifact may be defined as any recorded signal not originating in the brain. Artifacts are not always undesirable and physiological artifacts frequently provide clinically useful information, such as eye movements, including nystagmus, muscle artifact, tremors, body movements, respiration artifact and EKG. Artifacts do need to be properly identified. Sherlockian deduction and reasoning are helpful in interpreting EEGs; however, there are many artifacts that may appear identical or very similar to cerebral discharges. The ideal time to answer the question as to whether the activity being recorded is cerebral or artifact is during the recording, though this is often not possible with prolonged monitoring. The use of additional electrodes, the monitoring of the electrocardiogram (EKG), movements (e.g., body, tongue and eye), respiration and the temporary disconnection of other equipment may be needed to identify the non-cerebral origin of such activity. Failure to properly identify artifacts may have serious consequences. For example, misinterpretation of artifact as spikes or seizures may lead to misdiagnosis and inappropriate treatment. Artifacts may be divided into two groups: (1) physiological and (2) nonphysiological. At times, there is an overlap between these two categories. Physiological artifacts originate in the body but outside the brain. They include: The eyeballs and eyelids are factors in the production of eye movement artifact. The former acts as a dipole with the cornea being positive with respect to the retina. With eye closure or blinking, the eyeball rotates upward (Bell’s phenomenon) briefly, and the field of the strong corneal positivity involves the electrodes that are closest on the forehead (Fp1 and Fp2). Hence these electrodes become surface positive. The reverse is true with downward eye movement. Vertical eye movements are always maximal in the Fp1 and Fp2 electrodes and, therefore, the voltage gradient of this potential will often indicate its source. That is, if the activity is not greatest in Fp1 and Fp2 it cannot be a vertical eye movement. The converse, however, is not always true; some cerebral potentials may be maximal in the Fp1 and Fp2 electrodes. In order to resolve this difficulty, the simplest solution is to place additional electrodes below the eyes. Vertical eye movements will be out of phase when comparing superior (Fp1 and Fp2) and inferior orbital (LIO and RIO) electrodes referred to ipsilateral ear, while frontal slowing of cerebral origin will be in phase. There are some artifacts, such as a glossokinetic potentials (tongue movements), which can be mistaken for frontal or temporal slowing. This potential will be in phase between the superior and inferior orbital electrodes, and of higher voltage in the inferior orbital electrodes, which are closer to the tongue. Further monitoring of glossokinetic artifact is discussed later. Lateral eye movements are maximal in the F7 and F8 electrodes and are of opposite polarity (assuming conjugate gaze). Additional electrodes placed at the outer canthus of the eyes and referred to ipsilateral ear may help further delineate this activity since these electrodes (LOC, ROC) are closer to the source. It is standard to place eye leads by the right upper canthus and the left lower canthus (RUC, LLC); in that manner, all conjugate eye movements, horizontal or vertical, will be out of phase at those two diagonally-opposite locations (one is above the eyes, the other below; and one is to the left, the other the right). Eye movement artifacts are also discussed in Chapter 1: EEG basics, e.g., Figures and 1.12. An electroretinogram (ERG) – low voltage electrical signals generated in the retina – can be seen in some normal individuals if the Fp1 and Fp2 electrodes are used during photic stimulation. However, this artifact is most often seen in patients with cerebral death. In part this is due to the increased sensitivities employed in such recordings, as well as the absence of cerebral potentials. It does not invalidate the diagnosis of electrocerebral inactivity (ECI) or cerebral death, if other criteria are met. Eyelid flutter is also maximal in the Fp1 and Fp2 electrodes and often not well detected with infraorbital electrodes. At times, rapid eyelid flutter or nystagmus can give rise to confusion if not properly identified. In patients with a prosthetic eye or third nerve palsy, unilateral eye movement artifact may be mistaken for focal frontal delta activity. It is not unusual for a unilateral upgaze palsy (sometimes related to dysfunction of cranial nerve III) to be detected by the EEGer via asymmetric blink artifact. Muscle artifact is common in routine EEG recordings. Getting the patient more relaxed, if possible, is the best way to reduce this artifact. Occasionally the artifact can be unilateral, persist in sleep, and if severe, can obscure underlying cerebral potentials. Shivering artifact during therapeutic hypothermia is a common source of continuous prominent muscle artifact, even when shivering is not obvious on clinical observation (sometimes called ‘micro-shivering’). The use of the high-frequency filter (HFF) to reduce this activity can lead to difficulties in interpretation. For example, prominent muscle artifact may appear as beta activity when a HFF of 15 Hz is employed (avoid lowering the HFF that much!), or muscle spikes may be misinterpreted as epileptiform at this filter setting. The same is true of isolated muscle spikes seen with eye movements, particularly in the anterior temporal electrodes (F7 and F8), where lateral rectus spikes are commonly seen. There are commercially available modules, often as part of quantitative EEG measures, which have algorithms in place to assist with digital removal of muscle artifact. When it is critical to know if there is or is not cerebral activity underlying EMG artifact, such as with large myoclonic jerks in postanoxic coma, paralytics are occasionally recommended to allow definitive interpretation of the underlying EEG. Artifacts from tongue and mouth movements may be misinterpreted. The tongue acts as a dipole with the tip being surface negative and the base positive. Tongue movements in some patients will result in a prominent burst of slow waves, which may have a widespread distribution but are usually maximal anteriorly. If the technologist is aware of this, the artifact can easily reproduced by having the awake patient say ‘Tom Thumb’ or ‘lilt’. If that is not effective, performing vertical and/or lateral tongue movements may reproduce the pattern. This type of artifact can resemble a ‘projected’ rhythm or frontal intermittent rhythmic delta activity (FIRDA). Unfortunately, glossokinetic artifact can occur in patients who are confused and lethargic who may also have anterior frontal slowing that is cerebral. To help resolve this, additional electrodes can be placed above and below the mouth and slightly off-center in opposite directions on either side. A bipolar derivation from these perioral electrodes (above the mouth on one side and below the mouth on the other) will display this artifact clearly. Perspiration or sweat artifact is a long-duration potential (commonly lasting a few seconds) that usually involves several electrodes. Although the use of a short time constant (i.e., a higher low frequency filter [LFF] setting) will eliminate this artifact, it will also affect slow frequency cerebral activity. Therefore, this is not recommended, but rather one should try to decrease the sweating by cooling the head, applying alcohol or an antiperspirant to the area and having adequate air conditioning. EKG artifact is usually recognized by its regular, stereotyped appearance, and by comparing to a dedicated channel for chest recording. It is best seen on montages employing the ipsilateral ear as a reference and is less noticeable in bipolar montages or in referential recordings using Cz as a reference. Occasionally, irregular beats may be mistaken for cerebral potentials, or the coincidental coupling of an EKG artifact and a slow wave may give rise to the appearance of a spike-and-wave complex. Cardiac activity is best monitored by placing two additional electrodes on the chest and recording between them (to record cardiac potentials only, not a combination of cerebral and cardiac). The ear electrodes should not be used because they may record temporal spikes as well as the EKG. Occasionally EKG complexes may be mistaken for lambda waves. Pulse artifact can be mistaken for focal delta activity. This artifact is due to the movement of an electrode due to pulsation of an artery under it and is time-locked to the QRS complex. It usually occurs approximately 200 msec after the R wave (the estimated time taken for a systolic contraction of the heart to send a pulse of blood to a scalp vessel). Another way of thinking of it is that it usually aligns roughly with the T wave. A similar but more widespread artifact is seen in comatose patients, including cerebral death recordings, which is due to movement of electrodes, electrode wires, or the head from the mechanical ‘recoil’ effect of the beating heart (cardioballistic artifact). A pacemaker artifact is cardiac-related and results in ‘spikes’ that will be intermittent or regular, depending on the type of pacemaker. There are a wide variety of movement artifacts affecting the EEG. These include head tremors, body tremors that passively move the head, and body movements. The tremor can often be monitored with additional electrodes. Muscle and movement artifact are often prominent in patients having psychogenic nonepileptic seizures. Rhythmic activity in the occipital regions due to head movement can be eliminated by raising the patient’s head from the bed. Facial movement, such as during a focal motor seizure, hemifacial spasm, facial myokymia or facial synkinesias, can result in muscle artifact. Nonphysiological artifacts include: An electrode ‘pop’ is a common occurrence in any recording and will appear in all channels of the montage in which that electrode is present, but no others. An electrode pop occurs when there is an abrupt change in the electrical potential between the electrode and the scalp, which results in a sudden-onset, high voltage, usually positive (but occasionally negative) discharge, similar to the potential when a capacitor is discharged. An electrode artifact can occur intermittently or occasionally can be regular or even seemingly evolving (e.g., as an electrode is falling off the scalp). Findings limited to a single electrode should be viewed with suspicion and assumed to be artifact until proven otherwise. The use of an additional electrode placed next to the one in question can indicate whether the discharge is cerebral or artifact, if simple measures such as filling or replacing the electrode, etc., has not eliminated it. Electrodes with high impedance, usually due to poor electrical contact with the scalp, are more prone to show artifact, and unequal resistances frequently result in 60 Hz artifact. The electrode problem may be with the disc itself, but can also involve the wire, its connections or the jackbox into which the wire is plugged. Occasionally, one can get a ground electrode recording artifact without 60 Hz interference. The type of artifact seen will depend on the location of the ground electrode. In addition to ensuring low (<10 kohms) and relatively equal impedances (never >5 kohms difference between any two electrodes), accurate placement of electrodes is crucial to good recording technique, and a spurious asymmetry may result from inaccurate placement. Of the external sources of artifacts, 60 Hz is the most common. Although often encountered routinely in the laboratory, it can become a difficult problem in the ICU. Here, judicious disconnection of various equipment is often required to eliminate the artifact. Sixty Hz artifact is not necessarily bad since it often indicates electrode problems. Furthermore, the indiscriminate use of 60 Hz ‘notch’ filter will make recognition of electrode artifact more difficult; one should read without a notch filter on whenever possible. Artifacts may result from electrostatic interference. This often occurs with movement of persons in the vicinity of the patient and is most common in the ICU. Another artifact, which is rare, is due to an intravenous infusion and is also probably electrostatic in origin. There are a variety of instrumentation artifacts and include those due to an amplifier, settings (filters, sensitivity) and cables. During prolonged EEG monitoring when a technologist is not present throughout the recording, there are many other artifacts. The most troublesome are those that cause rhythmic artifact that can mimic seizures, such as patting and chest percussion by respiratory therapists or bed oscillators. Recording video is very helpful for the quick and accurate identification of these artifacts. Thus, there are multiple causes of artifact, both of a physiological and nonphysiological nature, and these need to be recognized by both the technologist and the electroencephalographer. They may not always be able to be eliminated, but they must be properly identified and proven, when in doubt, at the time of the recording.
7
Artifacts that can mimic seizures or other physiologic patterns
Figure list
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
