Rowan-06.

qxd

13/2/03

15:22

Page 75

6
TIPS ON READING AND
REPORTING THE EEG

The novice is confronted by a dizzying array of wiggly lines in a seemingly endless series of displays (or pages). At first the task seems impossible. How can I
ever master this arcane science (or art, or a combination of both)? The secret, as
in most complicated matters, is to break the EEG down to its essential elements.
At the end of the exercise, one reassembles the parts to make a cohesive, understandable whole. We think of it as providing the ordering clinician with a wellwrapped package, bow and all. After all, the whole point of EEG interpretation
is to help the clinician in his or her diagnostic quest.

ELEMENTS OF THE REPORT

One method of designing the EEG report breaks it down into five sections.
Each is important in constructing an accurate picture of the patients electrophysiological status.
1. Clinical Information. This section should include the reason for the EEG
request. The clinical information supplied by the clinician is often rudimentary; thus, the technician can be of assistance in obtaining additional
information from the patient and/or the patients chart. If the patient had
a recent seizure, say hours or a day before recording, it should be indicated
here. If the patient is taking medications that have effects on the brain,
they should be listed here. Examples include AEDs, psychotropic agents,
and sedatives.
2. Conditions of Recording. In this section, describe whether the EEG is
obtained under special circumstances such as after sleep deprivation or sedation. Then indicate the technicians estimate of the patients state awake,
drowsy, sleeping, lethargic, comatose, etc. Other observations by the technician should be noted: the patient might have been restless, tense, confused, moving throughout, chewing, etc. If the record was obtained at the

Rowan-06.qxd

76

13/2/03

15:22

Page 76

TIPS ON READING AND REPORTING THE EEG

bedside, or in an ICU or emergency room, it should be indicated here. All

these features set the stage for understanding the following section.
3. Factual Report. Here, cerebral and non-cerebral events are described. We like
to think that this section creates a word picture that more or less accurately
lays out the important findings. Bear in mind that the report may be read by
another electroencephalographer, and the record itself may be reviewed. Ask
yourself, would such a peer reviewer come to the same conclusion? Another
way to look at the issue is to ask, could I show a slide of this apparent spike
or sharp wave at a national meeting without fear of being hooted off the stage?
4. Impression. State in brief summary the essential findings. For example: This
is an abnormal EEG demonstrating, against a normal background, an active
spike focus in the anterior temporal region. Or: Abnormal EEG owing to
mild, diffuse slowing along with intermittent right temporal delta activity.
5. Clinical Correlation. This final section is perhaps the most important aspect
of the report. It is a truism that the average clinician does not read the detail
in the factual report, nor does he or she much care whether there is mu
rhythm over the left central region. The clinician wants diagnostic help, if
possible. The authors look at it this way: the factual report is important for
electroencephalographers; the impression and clinical correlation is of interest to clinicians. So, one offers as much assistance in this section as possible.
If the findings support a diagnosis of localization related epilepsy, say it here.
If the findings are consistent with a metabolic disorder, or an acute infarct,
then so indicate. This section is also the place to recommend further studies,
if indicated. For example, if there is a minor temporal abnormality in a
patient with probable complex partial seizures, you may suggest an EEG after
sleep deprivation. Or, you might recommend an ambulatory EEG.
Remember, serial studies are often helpful in certain situations (metabolic
disorders, for example). The clinician needs your guidance in ordering immediate additional testing and how he or she should proceed in the future.
A further consideration is what to do when the record is studded with artifacts. Adhering to the principle of aiding the clinician, one should try to make
some helpful statement even though, as a whole, the record is unsatisfactory for
accurate interpretation. A few pages of relatively quiescent recording may reveal
focal slow waves, or even a spike or two, giving some clue concerning the clinical problem. In these cases, one may recommend a repeat tracing when the
patient is less confused or less restless, or perhaps a tracing after mild sedation,
before a final impression is rendered.

HOW TO LOOK AT THE RECORD

In order to interpret the record properly, one must have clearly in mind
the elements of a normal EEG. This establishes a template, against which all

Rowan-06.qxd

13/2/03

15:22

Page 77

TIPS ON READING AND REPORTING THE EEG

77

deviations are to be compared. In the case of adults, recall that the normal
waking record contains an alpha rhythm (with some exceptions as outlined
previously), of maximum amplitude in the posterior quadrants. Beta activity is
present fronto-centrally, in greater or lesser amounts. A few theta frequencies are
acceptable if not lateralized. Little else is present. The response to hyperventilation may be marked in younger adults. Consistent lateralization is not permitted. A following (driving) response to intermittent photic stimulation is variable,
but there should be no activation of epileptiform activity. The elements of
drowsiness and sleep round out the template of the normal adult record.
It is sometimes useful to analyze carefully the first few interpretable pages or
10-second digital epochs. (Bear in mind that it may take several minutes before
the patient settles down and relaxes fully.) This exercise is often time well spent.
One may appreciate a hint of focality, or a hint of paroxysmal activity, or a possible asymmetry of background activity. In any event such analysis can set the
stage for more rapid analysis of subsequent recording, providing clues on what
to look for. Another hint that often pays dividends is to divide reading into
vertical and horizontal appraisals.
Horizontal reading appraises and compares channels on the left side of the
head with those on the right by scanning the record from left to right, choosing,
for example, the left and right temporal leads followed by the left and right
paracentral leads.
Vertical reading, scanning a particular segment of a second or so from top to
bottom, concentrates more on particular waveforms and their distributions. A
common problem with the beginner is that he or she becomes enmeshed in the
complicated polyrhythmicity of the EEG by reading only vertically. This results
in confusion and overreading. Overreading is a pitfall to be avoided (as, to some
extent, is underreading), but more on this later.
Reading horizontally reveals abnormalities that may not be at all evident
when reading vertically. In essence, horizontal reading reveals the broad picture
presented by the EEG. For example, this is the best way to determine alpha
asymmetry. One can easily determine alpha asymmetries, abundance, and irregularities. This is not evident with vertical reading. Unilateral or bilateral slowing
also becomes more evident during horizontal reading. Importantly, the appearance of a new, paroxysmal frequency, diagnostic of an electrographic seizure, may
not be evident with vertical reading. Many cases of recurrent electrographic
seizures have been missed due to this oversight.
Now, look for pages where the patient is in his or her most alert state. Here is
the point to analyze the alpha rhythm or the PDR. Is it well organized (that is,
nicely rhythmic, devoid of admixed slower frequencies), or is it irregular or poorly
persistent, or variable in frequency? Analysis of the PDR helps materially to determine laterality of any unilateral pathology. It is emphasized that if the patient is not
maximally alert, no definite determination can be made of a PDR asymmetry.
Now, evaluate the beta rhythm. Beta offers a few clues to the presence of
pathology. If asymmetric, it usually points to pathology on the side of diminished

Rowan-06.qxd

78

13/2/03

15:22

Page 78

TIPS ON READING AND REPORTING THE EEG

amplitude. If abundant, it may reveal a drug effect. In any case, it is an important

aspect of the basic or resting record.
The next task is to determine the presence of slow waves in the background.
The slowing may be diffuse that is, not lateralized and non-focal. Look for both
theta and delta frequencies. Is there a small amount of intermittent theta activity?
A great deal of theta? Are there mainly delta waves? A combination of the two? We
sometimes look at the broad picture and try to decide if it belongs in the theta box
or the delta box! Note that slowing may be more or less continuous, discontinuous, paroxysmal, rhythmic, or arrhythmic. There may be variable but inconsistent
lateralization.
One problem is to determine at the outset whether or not there is excess
slowing that is, how much is too much? This is a somewhat subjective call.
Major determinants include age and state, and the reader must constantly bear
these factors in mind. Look for pages in which the patient is fully awake. (If the
patient is drowsy throughout, this may not be possible.) Most electroencephalographers accept a small amount of theta in the background of adults as normal
(say up to 5%) mainly in the temporal regions.
Generally speaking, delta activity does not appear in the normal, waking adult
EEG. Perhaps a few random low-voltage delta waves might be acceptable. Note
that delta is common in the EEGs of children (see section on children). Also, in
the elderly, delta waves are commonly recorded in the temporal derivations,
indicative of some degree of cerebral pathology but not necessarily structural
disease demonstrable on an imaging study.
Now, describe any consistent focal slowing, theta or delta, that indicates localized cerebral pathology. The focality may be present in a setting of a slow disorganized background, or may stand out against a relatively normal background.
Finally, describe any bifrontal delta activity, its symmetry or asymmetry, rhythmicity, and amount.
After determining what the background of the record contains in terms of
slowing, evaluate the presence or absence of epileptiform activity. Describe the
type and location of any spikes, sharp waves, spike-wave complexes, or other
paroxysmal discharges (e.g. episodic sharp theta or delta frequencies, including
FIRDA). A statement concerning discharge abundance should be made. Focal
discharges such as spikes may be obvious, or hidden in the ongoing background.
Reliance on fairly strict definitions of epileptiform discharges is essential.
Sometimes, the record seems to contain sharp potentials everywhere! Careful
analysis will reveal that most of these putative sharp waves are simply the result
of a polyrhythmic background superimposition of various frequencies.
The alert recording having been assessed, now turn to the recording during
drowsiness. In most cases, subjects become drowsy during some part of the
record. Look for various drowsy patterns that are found in normal adult subjects.
These include, but are not limited to, slowing of the PDR, interruption of the
PDR by slower activities, anterior spread of a slowed PDR, generalized reduction
in amplitude, appearance or increase in diffuse theta activity, bifrontal delta

Rowan-06.qxd

13/2/03

15:22

Page 79

TIPS ON READING AND REPORTING THE EEG

79

activity in older subjects, and generalized rhythmic 4 to 5 Hz activity in children.

Note that subjects may drift in and out of drowsiness rapidly. This cannot be
determined by clinical observation. Also look for roving eye movements in the
anterior derivations; this is a good sign of drowsiness. There are relatively lowvoltage slow potentials, maximal in the frontal regions. Vertex sharp waves may
be seen during drowsiness and do not define entry into Stage II sleep. In addition, early sleep spindles may appear, usually less than 1 second in duration.
POSTS and K-complexes may also appear without full development. These
features, however, herald the approach of Stage II sleep.
Important concept: Recording of drowsiness is an essential element in EEG
diagnosis. Often, pathological findings occur mainly, or only, during this state.
Of particular importance is the precipitation, or exaggeration, of epileptiform
activity (spikes, spike-wave complexes also sharp waves if a broad definition is
accepted). Slow wave abnormalities are frequently exaggerated during drowsiness. Bifrontal rhythmic delta (FIRDA) may become more prominent during
this state. In any case, indicate the effect of drowsiness on the major findings.
After drowsiness, move to the main phenomena of sleep. Describe the features
of Stage II sleep (established sleep spindles, vertex sharp waves, K-complexes,
POSTS, and increased diffuse slowing in theta and delta ranges). At first, theta
predominates, followed by increasing amounts of delta. Note that vertex sharp
waves are not always sharp. On the other hand, they may be very sharp, and
resemble spike potentials. In addition, they may be isolated and sporadic, or
highly rhythmic.
Stage II sleep is followed by SWS. During this state the background is dominated by delta activity. At the same time, sleep spindles become less prominent
and may disappear. In general, SWS is infrequently recorded in adults during
routine EEG recording. On the other hand, SWS occurs frequently in young
children, especially after they have been sedated. Include in your description the
effect of sleep on any abnormalities previously noted or new abnormalities that
may appear.
You are now ready to dictate your report. Bear in mind always that you are
trying to convey what patterns you have seen, what they mean, and how you can
assist the clinician.

Rowan-06.qxd

13/2/03

15:22

Page 80