Electroencephalographic Artifacts

Wave Morphology

Certain artifacts in the EEG have characteristic shapes, such as eyeblink artifact, muscle artifact, and some electrode pops. Some of these artifacts have such a specific appearance that their identification essentially consists of a “sight diagnosis” or pattern recognition. The conclusion that some waves represent artifact may require a combination of both techniques, wave pattern recognition and analysis of wave topography. Examples of various specific types of artifact are shown later in this chapter.

“Biological Plausibility” Versus “Biological Implausibility”

Although some artifacts can be recognized on the basis of the shapes of their waves, the most powerful tool in distinguishing a true brain wave from an artifact is to establish its specific topography. Here, the major operating principle is that biological events recorded on the scalp tend to have a point of maximum voltage with the surrounding measured voltages diminishing with varying steepness from the point of maximum (see Figure 6-1 ). The voltage gradients seen with electrocerebral events are expected to have some degree of “smoothness.” In contrast, the electric fields of artifacts may have a patchy and unpredictable contour (compare to Figure 6-2 ).

A second principle is that the polarity of most discharges tends to be consistent. Most discharges, if negative at their maximum point, remain negative wherever they are detected on the scalp as shown in Figure 6-1 . In general, negative events tend to be negative everywhere, and positive events tend to be positive everywhere. Exceptions to this rule exist, including the occasional examples of discharges that manifest a so-called horizontal dipole (the example of the “horizontal” or tangential dipole is discussed in detail in Chapter 10 , “The EEG in Epilepsy”). Even in the small minority of epileptiform events that manifest a simultaneous combined positivity and negativity, those areas of opposite charge are usually segregated in a simple and orderly fashion rather than showing a pattern of several separated regions of positivity and negativity (see Figure 6-3 ).

Patterns with significant charge inconsistencies are also not expected from biological systems. The example of an apparent discharge with strong negativities in the left posterior quadrant and the right anterior quadrant but no measured voltage change in the intervening areas as depicted in Figure 6-4 is not likely of cerebral origin. Because this is not a “biological” pattern, there would be a strong suspicion of motion or some other type of artifact having affected two distant electrodes. This pattern and the pattern that was seen in Figure 6-2 may be produced by the haphazard jostling of electrodes during patient movement.

Whenever analyzing any wave on the EEG page, the reader should visualize the voltage topography of the discharge to confirm a distribution of charge that is potentially consistent with cerebral activity. Usually with a bit of practice this can be done quickly. Discharges with bizarre or unpredictable electrical fields likely represent artifact. The importance of localization skills cannot be overemphasized and it is for this reason that the techniques of localization discussed in Chapter 4 , “Electroencephalographic Localization,” should be mastered before moving on.

How do such “biologically implausible” artifacts arise in the first place? Most such artifacts are related to head motion. When the patient’s head moves, some combination of electrode wires is tugged on. The pattern of wires that is pulled with a head movement is, for the most part, unpredictable and may not follow any particular spatial pattern on the scalp. Disparate and distant deflections on the scalp, especially with varying polarities, are much more suggestive of motion artifact than of electrocerebral activity.

A type of motion artifact that might reliably follow a spatial pattern is head rolling or rocking. In such instances, the electrodes lying against the bed are most susceptible to motion artifact. For example, if the occipital portion of the patient’s head is in contact with the bed, head motion artifact will tend to be most dramatic in the occipital electrodes. Likewise, if the patient is lying on the left side of the head, the left temporal electrodes will be most prone to motion artifact. Artifact may appear in these or other electrodes with voluntary head movements, or even with the low-amplitude head movements associated with the patient’s respirations.

Some significant EEG diagnostic challenges occur when patients are referred for the question of whether sudden movements are epileptic. The diagnostic problem is made more difficult in that each movement may generate a motion artifact causing the appearance that each is correlated with an EEG burst. There would be no difficulty if the movement in question were, for instance, a nonepileptic twitching of the hand because a hand movement would not cause motion artifact in distant scalp electrodes. The hand movement would occur during the study, and there would be no EEG change, strongly suggesting that the hand movements were nonepileptic. If the movement involves the head, however, there may be a simultaneous burst in the EEG, creating some confusion as to whether the movement caused the wave (and is an artifact) or the wave caused the movement (and is an epileptiform discharge). The approach to this problem takes into account that when the head moves, is turned on the pillow, or is subject to some other type of external movement, the pattern of electrodes that are disturbed tends to be more random. Artifacts caused by electrode motion or head motion tend to show a nonregular distribution of inconsistent voltages like those shown in Figures 6-2 and 6-4 . In comparison, the patterns associated with electrocerebral activity are expected to show more consistent polarities and a voltage gradient that tapers with distance from the point of maximum.

A second approach to this type of problem is to analyze the way in which the EEG discharge is time-locked to the movement. When a movement is caused by an EEG burst (e.g., epileptic myoclonus), there is always a slight time lag between the EEG burst and the movement because it takes some amount of time for the cortical signal to propagate through the nervous system to the muscles. This latency between the burst and the movement may be confirmed either by a synchronized video recording of the patient or an electrode placed on an involved limb. With use of the latter, the EEG burst appears first, followed by the muscle burst in the limb electrode. If an apparent EEG burst is really an example of motion artifact, the two are generally seen to occur on the video simultaneously, or the body movement may even start before the EEG deflection.

Artifacts Associated With Eye Movements

Eyeblink Artifact

Electroencephalographers know that when individuals close their eyes, the globes of the eyes deviate upward. This comes as a surprise to some because when we watch a person casually blink or close the eyes, this upward movement of the globes is hidden from view by the closed eyelids. It is only in special situations that this reflex upward eye deviation with eye closure, termed Bell’s phenomenon, is readily observable. Bell’s phenomenon becomes strikingly evident in the case of individuals who suffer from Bell’s palsy. In Bell’s palsy, there is a paralysis of the facial nerve (cranial nerve VII), one of whose functions is to close the eyes. The nerves that move the globe of the eye (cranial nerves III, IV, and VI) are unaffected in patients with Bell’s palsy. Therefore, Bell’s palsy patients have normal eye movement but cannot close the affected eyelid. Thus, the facial nerve paralysis caused by Bell’s palsy literally uncovers Bell’s phenomenon—we can see what the globe of the eye is doing during intended eyelid closure. Bell’s palsy patients have normal eye closure on the unaffected side, but when blinking or closing the eye on the paralyzed side, the globe of the eye can be seen to deviate upward, unhidden by the eyelid. Upward deviation of the globes may also be seen when a person’s eyelids are forcibly pried open during attempted voluntary eye closure. Finally, in some individuals who are in the process of falling asleep, especially babies, the eyelids may lower to a “half-mast” position but do not completely close. Because of a partial Bell’s phenomenon associated with the partial eye closure, only the whites of the eyes can be seen in the half-open palpebral fissures.

The type of artifact that an upward movement of the globes might cause during eye blinking or eye closure is not intuitively obvious without knowing that the globe of the eye has a particular distribution of charge. As it happens, the cornea (the front of the globe of the eye) carries a net positive charge and the posterior pole of the globe carries a net negative charge, forming a dipole. Because of the presence of this positive charge on the cornea, the bobbing upward of the eye with eye closure is easily detected by EEG electrodes. Figure 6-5 illustrates how the EEG electrodes closest to the eye “perceive” an eyeblink. With upward eye deviation, the electrodes closest to the globes, Fp1 and Fp2, perceive the strong positivity of the corneal surfaces. The F3 and F4 electrodes, which are immediately posterior to the Fp1 and Fp2 electrodes, also pick up some of this positivity, but to a much lesser degree than the Fp1 and Fp2 electrodes. With Fp1 being more positive than F3 and Fp2 being more positive than F4, the Fp1-F3 and Fp2-F4 channels show the characteristic sharp, downward-sweeping waveforms of eyeblink artifact (see Figure 6-6 ). A similar type of artifact is seen with eye fluttering (see Figure 6-7 ).

A useful general rule regarding eyeblink artifact is that it should not be detectable in the electrodes of the posterior half of the head; movements of the globes of the eye should not be perceived by the parietal or occipital electrodes because they are too distant. Eyeblink artifact is only occasionally picked up by the central electrodes, C3 and C4, and when it is it should be of low voltage (see Figure 6-8 ).

Lateral Eye Movement Artifact

An extension of this same phenomenon involving the motion of the positive charge on the anterior aspect of the globe is used to understand the appearance of lateral eye movement artifact. In the case of lateral eye movements, the anterior temporal electrodes, F7 and F8, are the electrode contacts best placed to perceive the positivity of the cornea as the eyes move from side to side in the horizontal plane. For example, when a patient gazes to the left, the F7 electrode picks up a net positivity. Simultaneously, as the eyes gaze to the left, the F8 electrode picks up a net negativity from the posterior pole of the globe (see Figure 6-9 ). This results in a characteristic pattern seen with a simultaneous “positive phase reversal” in the anterior temporal electrode on the side to which the eyes have turned and a “negative phase reversal” in the anterior temporal electrode on the side away from which the eyes have turned. In the case of more rapid voluntary saccades or REM sleep, these lateral eye movement artifacts can have a squared-off appearance at wave onset. During drowsiness, slow roving lateral eye movements may be seen in which the artifact shows a more rounded appearance (see Figure 6-10 ). Therefore, identification of both eyeblink and lateral eye movement artifact requires a combination of wave pattern recognition and localization of the electrical events to their characteristic locations.