Artifact Recognition and Technical Pitfalls

Artifact Recognition and Technical Pitfalls

Thoru Yamada

Elizabeth Meng

Artifact Recognition

Artifacts are recorded activities that originate somewhere other than the area of interest. When recording an ECG (electrocardiogram), anything recorded which does not originate in the heart is considered an artifact. When recording EEG, anything recorded that is not cerebral in origin is considered an artifact. No EEG is free of artifacts. Identifying artifacts correctly is an essential role of an expert in EEG. This applies to both EEG technologists and electroencephalographers. Identifying, documenting, and eliminating artifacts are major roles of an EEG technologist. Some high-amplitude artifacts can totally obscure EEG activity. Some are of low amplitude and can subtly minimize or distort the cerebral activity. The artifacts are, however, not always useless. Some artifacts provide crucial information for an appropriate interpretation. Correlating artifacts with an ongoing EEG is an integral part of a skilled EEG interpretation. Artifacts can be grouped by origin such as:

  • Artifacts of physiological origins

  • Artifacts associated with body or head movements

  • Artifacts of nonphysiological (electrical) origins


Physiological artifacts originate from the body itself. The contamination of physiological artifacts is inevitable in any EEG, though artifacts can be modified or minimized by appropriate technical adjustments.

Myogenic Artifacts (EMG; Electromyography)

EMG artifacts arise from nearby muscles introducing “muscle artifacts” mostly in the temporal and frontal electrodes. In spite of the technologist’s efforts to relax the patient and reduce muscle artifacts, some patients, especially the elderly or uncooperative ones, may continue to show tonic muscle artifacts. The frequency of EMG artifacts varies from patient to patient.

When EMG is excessive and unrelenting, it can be minimized by lowering the high filter from 70 to 35 or 15 Hz. When this is necessary, caution should be exercised because filtered muscle can resemble beta activity (see Fig. 15-39A and B). Also, filtered muscle artifacts could mimic “spike” discharges (see Fig. 15-40A and B). EMG artifacts are usually prominent in the ear or mastoid electrodes because of the proximity to the temporalis muscles. If the contamination of artifacts is excessive, the reference can be switched to the vertex (Cz) where muscle artifacts are minimal (Fig. 15-1A and B).

Tonic muscle artifacts, on the other hand, provide a useful clue to determine the patient’s level of consciousness. Tonic muscle artifacts usually diminish during drowsiness. An awake or drowsy-appearing EEG coupled with an absence of tonic muscle artifact raises the possibility of rapid eye movements (REMs) sleep (see Fig. 7-45F). REM, scanning eye movement while eyes are open, or nystagmus causes a lateral rectus muscle twitch artifact (this is caused by activation of cranial nerve VI), which is recorded at F7 or F8 electrodes (see Figs. 15-9 and 15-10).

Muscle artifact may appear as a single muscle “twitch.” This discharge is usually of a much shorter duration than an epileptiform spike. When they occur in sleep, however, “muscle spikes” tend to have a longer duration, and sometimes resemble the spikes of cerebral origin. Unusually, narrow and closed field distribution or random and scattered occurrence of these muscle spikes aids to distinguish them from epileptiform spikes (Fig. 15-2A). Repetitive muscle twitch artifacts resemble repetitive spikes or ictal (seizure) discharges or may obscure the background activity (Fig. 15-2B).

In comatose patients with decerebrate rigidity, shivering, or myoclonic twitches, pharmacologic paralysis may be necessary to distinguish the cerebral activity from artifacts (see Fig. 10-35A and B). Drugs like rocuronium, pancuronium, or succinylcholine can be used only when the patient is intubated, mechanically ventilated, and with an order from a physician (see Video 13-7). Swallowing or chewing results in a distinctive and diffuse crescendo-decrescendo type of muscle artifacts (see Fig. 15-18A and B).

Eye Movement Artifact

The eyeball can be regarded as a dipole with positivity toward the cornea and negativity toward the retina. When the eyeballs are in a fixed position, the dipole does not yield any potential change in the EEG. But, when the eyeball moves, this moving dipole generates a large and slow AC potential, detected by electrodes near the eyeballs. When the eyes close or blink, both eyeballs move in a conjugate upward direction (the Bell phenomena). This results in a positive deflection, which is maximal at Fp1 and Fp2 (Fig. 15-3, Video 15-1A, see also Fig. 8-16A and B).

Conversely, when the eyes open, a downward eye movement causes a negative potential at Fp1 and Fp2. Horizontal eye movements are reflected maximally at F7 and F8 with these two electrodes being charged in opposite polarities. Looking to the left causes positivity at F7 and negativity at F8; the reverse is true when looking to the right (Video 15-1B, see also Fig. 15-3).

FIGURE 15-1 | EMG artifact arising from the left ear (A1) (A). Because of close proximity of the ear to the temporalis muscle, ear reference recording is often contaminated by EMG artifact. This can be minimized with the use of Cz reference (B). (A and B are the same EEG samples.)

FIGURE 15-2 | Examples of “muscle spike” artifact recorded from frontal electrodes (A). The morphology of this discharge resembles a real “spike,” but the narrow field distribution of this discharge (restricted to Fp1 and Fp2) revealed on a double banana montage supports this as an artifact, not a cerebral potential. Another example is repetitive “muscle spikes” resembling ictal (seizure) discharges (B).

FIGURE 15-2 | (Continued)

FIGURE 15-3 | Eye movement artifacts of various directions. Eye opening shows a negative deflection at Fp1 and Fp2 (A). Eye closing or blink shows a positive deflection at Fp1 and Fp2 (B). Horizontal eye movements show opposite polarity between F7 and F8; with right horizontal movement, F8 becomes positive (shown by circles) and F7 becomes negative (shown by rectangular box) (C). The deflections are reversed for left horizontal eye movements (D).

In order to determine with certainty if a discharge represents eye movement or real cerebral activity, it may be necessary to use additional electrodes called eye monitors. In monitoring eye movements using two channels on each side, one electrode is placed just above the eye (Fp1 and Fp2 can be used), and one just below the eye (infraorbital: IO) on each side. Each is referred to the ipsilateral ear reference. Using this montage, the relationship between Fp and IO is out of phase for vertical eye movement and in phase for cerebral activity (see Fig. 15-4A and B, see also Fig. 8-16A).

An alternative derivation utilizes electrodes placed above the outer canthus of the left eye (LOCa) and below the outer can thus of the right eye (ROCb) or vice versa. Recording between LOCa-A1 and ROCb-A2, both vertical and horizontal eye movements, will show an out-of-phase deflection (see Fig. 15-4A and B, see also Fig. 8-16B) (see also Video 15-1A and B).

Without eye monitors, it can be difficult to differentiate bifrontal delta activity from vertical eye movement artifact. In general, the potential gradient of an eye movement from frontal to posterior electrodes is steep and rapidly dissipates in the posterior electrodes. This gradient is more gradual for frontal delta activity of cerebral origin (Fig. 15-5).

Some subjects have subtle and rapid eyelid fluttering accompanying rhythmic artifacts. This causes rhythmic theta or alpha range activity at Fp1 and Fp2 electrodes called eye flutter artifacts (Fig. 15-6A) (Video 15-2A). Some can even control the frequency of eye flutter (Video 15-2B). Eye flutter artifacts may also be induced by repetitive photic stimulation (Fig. 15-6B).

FIGURE 15-4 | Recording of eye movements with eye monitor electrodes placed at the left and right infraorbital region (LIO/RIO at channels 3/4), just above the left outer canthus, and just below the right outer canthus (LOCa/ROCb at channels 5/6). Vertical eye movements recorded from infraorbital electrodes are out of phase with the frontopolar electrodes. Lateral canthus electrodes are out of phase with each other for either eye-opening or eye-closing movements (A) and also for horizontal eye movement (B). Note that the diffuse but bifrontal dominant inphase delta activities at all electrodes are consistent either with glossokinetic or frontal dominant delta activity potential (shown by box); compare with Figure 15-11.

Asymmetric eye movement artifacts can mimic focal delta activity. This can occur when the patient has a prosthetic eye or diseased eye ball (Fig. 15-7). A severely diseased eyeball with a loss of normal ocular potential could also result in asymmetric eye movement artifacts despite the eyes moving in a normal and conjugate manner. Of course, asymmetric eye movement artifacts can be expected if there is asymmetrical electrode placement between homologous electrodes. An astute technologist should notice the asymmetries and clarify the reason for asymmetric eye movement potentials.

Another important artifact to be recognized is slow-drifting (horizontal) eye movements. This signifies that the patient is becoming drowsy and should not be disturbed if a sleep recording is desired (Fig. 15-8). Rapid horizontal eye movements are often accompanied by lateral rectus muscle twitch artifact, which can be detected at F7 or F8 electrodes. Because a lateral rectus muscle twitch causes the eye to move toward the contracted muscle, the twitch artifact is always followed by a deflection of positive polarity. This is best seen in REM sleep (see Fig. 7-45F) or in a patient who has nystagmus (Fig. 15-9).

Horizontal eye movements can be recognized because of the opposite polarities between F7 and F8 electrodes. Figure 15-10 shows unusual horizontal eye movement of rhythmic character (opsoclonus) resembling frontal delta activity (compare with Figs. 15-8 and 15-9).

FIGURE 15-4 | (Continued)

FIGURE 15-5 | Different field distribution between eye movement and frontal dominant delta activity. The potential fields of eye movement (blink artifacts) are relatively restricted to Fp1 and Fp2 electrodes (shown by rectangular boxes), whereas frontal delta activity (FIRDA/frontally dominant GPD* in this case) shows much wider distribution spreading posteriorly (shown by circles).

FIGURE 15-6 | An example of eye flutter artifact. Rapid eyelid movements associated with fast blinking called “eye flutter” show rhythmic alpha or theta range activity localized to Fp1 and Fp2 electrodes (shown by rectangular box) while eyes are closed (A). Eye flutter artifacts can be differentiated from cerebral activity by their restricted distribution limited to Fp1 and Fp2 electrodes. In order to verify this, eye monitor electrodes would be helpful (see Fig. 15-4A). The same patient had similar eye movements induced by photic stimulation with the frequency time locked with photic flashes (B).

FIGURE 15-7 | A patient with prosthetic eye on the right is shown by the absence of eye movement (blink) artifacts at Fp2 (compare the patterns, shown in rectangular box and in oval circle).

FIGURE 15-8 | Slow-drifting eye movements during approaching drowsiness. Note opposite polarities between F7 and F8 electrodes concomitantly (positivity is marked by asterisk and negativity is marked by #).

FIGURE 15-9 | An example of rapid horizontal eye movements (nystagmus). Note repetitive lateral rectus muscle twitches mostly at F7 electrodes (examples are indicated by rectangular boxes), followed by positive deflection (shown by oval circle) with concomitant negative deflection at F8 electrode.

FIGURE 15-10 | Unusually rhythmic eye movements (opsoclonus) artifacts seen in a patient with acute hemorrhagic stroke in the thalamus extending to the brainstem. This is not frontal delta activity but horizontal eye movements. The eye movements are verified by opposite polarity between F7 and F8 electrodes. Note positive deflection F7 (A) and F8 (B) indicating eye movement to the left and to the right, respectively.

Glossokinetic Artifact

Another artifact which resembles frontal delta activity or eye movement artifacts is glossokinetic potential associated with tongue movement. The tongue, like the eye, is electrically charged with negativity at the tip and positivity at the root. When the tongue moves, especially when touching the roof of the pharynx, the change in the electrical field spreads to the scalp. This causes single or rhythmic diffuse delta waves, especially prominent in the frontal region resembling vertical eye movement artifact (Fig. 15-11). This glossokinetic potential can be reproduced by asking the patient to say words including an “L” sound such as “lilt” (Video 15-3). Unlike the eye movement potentials, the contamination of the glossokinetic potential to the scalp electrodes varies considerably from one person to another; the same tongue movement brings out large artifacts in some but not in others.

The more difficult differentiation is between glossokinetic potential and frontal delta activity. The infraorbital electrode helps in the differentiation. Both glossokinetic potential and frontal delta activity show an inphase deflection between Fp (frontopolar) and IO (infraorbital) electrodes, but the amplitude of a glossokinetic potential is larger at the IO than the Fp electrode, while frontal activity is larger at the Fp than the IO electrode (Fig. 15-12, see also Fig. 8-16) (Video 15-4). Without eye monitor electrodes, vertical eye movements and glossokinetic potentials can be differentiated by the wider distribution of the latter than the former (Fig. 15-13).

FIGURE 15-11 | The relationship between blink artifacts and glossokinetic potentials. The blink artifacts are out of phase between Fp and infraorbital (IO) electrodes, whereas the glossokinetic potentials show inphase between the two with the greater amplitude at the IO electrodes than Fp electrodes. (This distinguishes between glossokinetic potential and frontal delta slow waves; see also Fig. 15-12A and B, also Video 15-3.)


Electroretinogram (ERG) is seen as a response to photic stimulation and is conventionally recorded from a contact lens electrode placed directly over the eyeball. In some subjects, however, the ERG can be recorded from Fp1 and Fp2 electrodes. The potential consists of two major components, “a” and “b” waves appearing as a small sharp and wave complex (Fig. 15-14A). The response is time locked to the flicker frequency. ERG is often seen in electrocerebral silence (ECS) recordings because of the high-amplification recording with absence of interfering EEG activity and still remaining function of the retina. These normal physiological responses from the retina must be distinguished from the nonphysiological artifact in which the electrode reacts to a light source. This can be differentiated by delivering high-frequency flashes (>30 Hz). The ERG cannot react to high-frequency flashes (Fig. 15-14B), but electrode artifact continues without diminishing amplitude. Alternatively, the electrode in question may be covered by an opaque material. ERG persists but electrode artifact disappears.

Cardiac Artifacts

ELECTROCARDIOGRAM ARTIFACT. The artifacts from electrocardiogram (ECG) can usually be identified easily because of their regular form and repetition. If, however,

ECG rhythm is irregular or has an abnormal waveform, the contamination of ECG artifact onto the EEG could mimic sharp or spike discharges. For this reason, ECG monitoring should be done routinely.

FIGURE 15-12 | An example for differentiation of frontal delta activity, eye movement artifacts, and glossokinetic potential using eye monitor derivations. The eye movement (vertical) artifacts show out-of-phase relationship between Fp and IO, but frontal delta activity shows inphase relationship between Fp and IO electrodes and also between LOCa and ROCb electrodes (A). The glossokinetic potential shows inphase relationship between IO and Fp electrodes (B) but differs from frontal delta by greater amplitude of activity at IO than Fp electrode. It is not possible to differentiate frontal delta activity from eye movement or glossokinetic artifacts if examining only Fp electrodes because of similarity of these three activities.

FIGURE 15-13 | An example of glossokinetic potentials produced by tongue movements when the patient was verbally counting. Note the difference in distribution between glossokinetic potentials and eye movement (blink) artifacts; glossokinetic potentials are more widely distributed with a slower potential gradient from the anterior to posterior head region (shown in rectangular box) as compared to the eye movement artifacts, which are more restricted to Fp1/Fp2 with a steep gradient from the anterior to posterior (shown in circle).

FIGURE 15-14 | An example of ERG (electroretinogram) seen in Fp1 and Fp2, produced by repetitive photic flashes (A). Each ERG, time locked with each flash, consists of a small spike-wave complex referred to as “a” and “b” waves (one ERG is enlarged shown by the box). ERG responds to the progressively changing flash frequency but becomes smaller with faster frequencies and fades out at around 18 Hz (B) (shown by rectangular box).

Only gold members can continue reading. Log In or Register to continue

Nov 14, 2018 | Posted by in NEUROLOGY | Comments Off on Artifact Recognition and Technical Pitfalls
Premium Wordpress Themes by UFO Themes