CONTENTS

• When & how long to monitor EEG:
  • Indications for EEG monitoring in the ICU
  • Duration of EEG monitoring
• Basics
  • EEG leads
  • Background pattern
  • Reactivity
  • Paralysis to eliminate artefacts
• EEG in encephalopathy/coma
  • Beta coma
  • Alpha coma
  • Theta coma
  • Delta coma
  • Spindle coma
• Seizures
  • Epileptiform discharges
  • Seizures
  • BIRDs – Brief Potentially Ictal Rhythmic Discharges 🐥
• Ictal-Interictal continuum
  • GPDs – Generalized Periodic Discharges
  • LPDs – Lateralized Periodic Discharges
  • LRDA – Lateralized Rhythmic Delta Activity
  • BIPDs – Bilateral Independent Periodic Discharges
  • SIRPIDs – Stimulus-Induced Rhythmic, Periodic, or Ictal Discharges
• Other abnormalities
  • GRDA – Generalized Rhythmic Delta Activity
  • Extreme Delta Brush
  • CAPE – Cyclic Alternating Pattern of Encephalopathy
• Podcast
• Questions & discussion
• Pitfalls

---

indications for EEG monitoring in the ICU

(back to contents)

---

There is a broad variety of indications for EEG monitoring in the ICU (e.g., for any patient in whom there is a concern for nonconvulsive status epilepticus – as discussed further here). Thus, the decision to pursue EEG monitoring is a complex one, which will often involve neurology consultation. Below are some rough guidelines (which are not intended to replace expert judgement).

general rationale

• Seizures or nonconvulsive status epilepticus are commonly detected among ICU patients (including “medical” or “surgical” ICU patients).(LaRoche 2018) Recognition of seizures is generally important for two reasons:
  • (1) Seizures may impair consciousness, thereby clouding neuroprognostication and delaying recovery/rehabilitation.
  • (2) Nonconvulsive status epilepticus may directly cause neurological injury (constituting one mechanism of secondary injury that can exacerbate critical neurologic illnesses).
• EEG monitoring is safe, so the main drawback is that it incurs some cost. However, by facilitating more rapid recovery and extubation, EEG monitoring may be cost-neutral or cost-saving.

general indications for prolonged EEG monitoring (34618762)

• Persistently abnormal mental status after generalized convulsive status epilepticus or clinically evident seizure(s).
  • A postictal period is expected following seizure(s), but this should usually clear within an hour.(LaRoche 2018) Persistent alteration of mental status suggests the possibility of ongoing seizure activity.
  • Nonconvulsive status epilepticus is commonly encountered following convulsive status epilepticus. However, nonconvulsive status epilepticus can occur following only a single convulsive seizure.
• Assessment of treatment efficacy in patients with known status epilepticus (e.g., titration of antiepileptic infusions to target absence of seizures or burst-suppression).
• A screening (“spot”) EEG has revealed patterns along the ictal-interictal continuum, which increase the likelihood of subsequent seizures.
• Paroxysmal clinical events suspected to be possible seizures.
• Patients at risk for seizures which may be masked by the requirement for pharmacologic paralysis.
• Fluctuating mental status or unexplained alteration of mental status.
  • Nonconvulsive status epilepticus or subclinical seizures occur in ~15% of critically ill patients with encephalopathy, even in the absence of acute brain injury (i.e., medical or surgical ICU patients).(34798964)
  • This includes patients with acute supratentorial brain injury whose mental status is worse than would be expected.
• Comatose patients following cardiac arrest. 📖
• Subarachnoid hemorrhage patients, to surveil for delayed cerebral ischemia (DCI).

---

duration of EEG monitoring

(back to contents)

---

sensitivity of EEG for seizure increases with longer duration of monitoring

• A single “spot” EEG recording over an hour may have only ~50% sensitivity for detecting seizure, as compared to ongoing continuous EEG monitoring. Increasing monitoring for 24-48 hours further increases sensitivity.
• Beyond a certain duration, the value of ongoing EEG monitoring becomes murky. For example, if the patient is having a subclinical seizure every 30 hours, does that truly matter? Would treating such an abnormality lead to more benefit than harm? This is unknown in the context of critical illness (wherein seizures may often improve over time merely due to treatment of the underlying disorder).

continuous versus spot EEG (aka “screening EEG”)

• Continuous EEG has greater sensitivity for seizures, but this comes at increased cost. Additionally, a hospital has only a limited number of EEG machines available. Consequently, it is important to determine which patients truly require continuous EEG monitoring.
• As a rough rule of thumb: If the following criteria are met, a short EEG recording may be sufficient to exclude seizures with reasonable sensitivity:(31886441)
  • (1) Lack of prior seizure history.
  • (2) No ongoing coma.
  • (3) EEG recording reveals no worrisome findings (e.g., lateralized periodic or rhythmic delta activity).
• The 2HELPS2B algorithm (below) provides a more sophisticated and evidence-based approach for determining which patients require extended EEG monitoring.

2HELPS2B risk score for predicting electrographic seizures

• 2HELPS2B is a validated algorithm that predicts the risk of seizures based on clinical information combined with an initial 1-hour screening EEG. This helps determine how long it is appropriate to monitor the patient on EEG (e.g., a lower score would support more prompt discontinuation of EEG monitoring).
• Components of algorithm:
  • 1 point each:
    • Frequency >2 Hz for any periodic or rhythmic EEG pattern (except for GRDA).
    • Epileptiform discharges (sporadic).
    • Lateralized pattern (LRDA, LPD, BIPD).
    • “Plus” features (including superimposed rhythmic, fast, or sharp activity).
    • Seizure history, or recent events suspicious for clinical seizures.
  • 2 points: BIRDS 🐥. 📖
• Risk of seizure:
  • 0 points: ~3-4%
  • 1 point: ~12%
  • 2 points: ~27%
  • 3 points: ~50%
  • 4 points: ~73%
  • 5 or more points: ~88%
• Implications of the duration of EEG monitoring (noting that clinical judgement is also involved):
  • 0 points: Screening EEG alone may be adequate.
  • 1 point: 12 hours of EEG screening may exclude seizure with ~95% sensitivity.
  • ≧2 points: 24 hours of EEG screening or more may be needed if the goal is detection of >95% of all seizures.
• Limitations:
  • (1) This is all based on an initial EEG recording of one hour (which may be longer than the duration of a spot EEG at many centers).
  • (2) This may be confounded by sedative/anesthetic agents on board during the screening EEG.
  • (3) The “prescription” for duration of EEG monitoring doesn't take into account the findings during ongoing EEG monitoring. For example, imagine that a patient has a 2HELPS2B score of “1” and is subsequently found to have BIRDs during their 12-hour continuous EEG – that patient would undoubtedly warrant ongoing monitoring for >12 hours.

---

EEG leads

(back to contents)

---

standard international 10-20 EEG lead placement system:

• Laterality:
  • Odd numbers = left side.
  • Even numbers = right side.
  • “z” = midline.
• Lobes:
  • F = Frontal.
  • T = Temporal.
  • C = Central.
  • P = Parietal.
  • O = Occipital.

---

background pattern

(back to contents)

---

different frequencies

• beta (>13 Hz) 📖
  • Beta activity is normally seen in frontocentral leads when awake.
  • Beta activity may be accentuated by light sedation (e.g., benzodiazepines or propofol).
  • Focal attenuation of beta activity may result from a cortical lesion (e.g., tumor, abscess, stroke) or from an intervening fluid collection (e.g., epidural or subdural hematoma).
• alpha (8-13 Hz) 📖
  • Normal patients with closed eyes should have a background alpha rhythm which is predominant in the posterior leads.
  • Eye opening or mental activity should attenuate the alpha rhythm.(33456874)
• theta (5-7 Hz) 📖
  • May be seen in drowsiness or encephalopathy.
• delta (<5 Hz) 📖
  • May be seen in sleep, encephalopathy, or structural lesions (focal delta).(Albin 2022)

continuity

• Continuous EEG is a normal finding.
• Nearly continuous: <10% is suppressed (<10 uV) or attenuated (>10 uV but <50% background amplitude).
• Discontinuous (10-49% of the record is attenuated or suppressed).
• Burst suppression or burst attenuation: >50% is suppressed or attenuated.
  • Further discussion of burst suppression here: 📖
• Suppression or attenuation: The entire EEG is suppressed (<10 uV).
  • Further discussion: 📖
• Electrocerebral inactivity: The entire EEG is totally suppressed (<2 uV).
  • Further discussion: 📖

findings reflecting sleep

• K complexes and sleep spindles are findings which may be seen during normal Stage-II non-REM sleep.
• These findings suggest normal sleep. They are positive prognostic signs.
• Features of sleep may also be seen in spindle coma. 📖

focal slowing

• This may reflect an underlying structural or functional abnormality.
• Intermittent focal slowing suggests seizure (as a postictal finding), migraine, or transient ischemic attack (TIA).
• Continuous focal slowing suggests a structural lesion (e.g., ischemic stroke, tumor, abscess, intracranial hemorrhage).

---

EEG reactivity

(back to contents)

---

EEG reactivity

• Reactivity refers to any clear change in EEG pattern in response to stimulation. This is generally a positive prognostic sign.
• Historically, the value of EEG reactivity has been limited by poor inter-rater reliability. This is due to:
  • Lack of standardization of the stimuli applied to the patient (e.g., some practitioners may administer more aggressive noxious stimulation than others).
  • Lack of standardization regarding how much of a change in the EEG constitutes “positive reactivity.”
  • Inadequate repetition of stimulus application.

consensus statement on determining EEG reactivity (30056156)

• Do not stimulate the patient right before reactivity testing.
• Stimuli should be applied in an escalating fashion (from less noxious to more noxious).
  • At a minimum, this should include clapping, calling out the patient's name, and nail bed pressure.
  • Additional stimuli that may be used include sternal rub, trapezius muscle squeeze, and passive eye opening. A central noxious stimulus such as sternal rub may be especially important among patients who might have peripheral neuropathy (e.g., due to diabetes).
  • 💡 Multiple types of stimuli should be applied, as patients may respond to some more than others.(28187794)
• Stimuli should be applied with a minimum duration of 5 seconds. Each stimulus should be repeated at least three times.
• The EEG is reactive when the change in response to a stimulus is reproducible (including at least two out of three stimulations).
• The following EEG changes don't qualify as reactivity:
  • Eye-blink artefact.
  • Motor artefact (including that stimulated by spinal reflexes).
  • Electrographic seizure induction.

clinical utilization of EEG reactivity?

• If the patient displays unequivocal EEG reactivity, that is a positive prognostic sign. However, care should be taken to avoid confusing EEG reactivity with the presence of EEG artefacts (e.g., motor artefact due to a spinal muscle reflex).
• If there is a lack of EEG reactivity, this is nonspecific. In particular, absence of reactivity doesn't necessarily predict a poor prognosis.
• ⚠️ Following anoxic brain injury, it's doubtful whether assessment of EEG reactivity provides independent information above and beyond other features of the EEG (such as background rhythm). Evidence regarding reactivity is much less robust than evidence surrounding other aspects of EEG analysis.

---

paralysis to eliminate EEG artefacts

(back to contents)

---

Muscle artefact may render EEG interpretation difficult or impossible (e.g., due to myoclonus or shivering). In situations where EEG information is critically important, it may be helpful to temporarily paralyze patients to obtain high-quality EEG data. However, this must be done carefully to avoid paralyzing a patient who could be awake.

prerequisites for paralyzing a patient to obtain better EEG data

• (1) The patient must be intubated and sufficiently stable to tolerate paralysis from a cardiopulmonary standpoint:
  • Negative respiratory efforts may increase venous return to the heart. In some exceptionally unstable patients, paralysis may impair cardiac output.
  • Patients with profound hypoxemia on some modes of ventilation (e.g., airway pressure release ventilation) may experience hypoxemia due to paralysis.
  • (The vast majority of intubated patients will tolerate transient paralysis without any cardiopulmonary problems.)
• (2) The patient must be clinically comatose (e.g., unresponsive to painful stimuli).
• (3) The patient should be on a stable regimen of sedating agents:
  • It is usually desirable to hold any sedating agents prior to neurological examination. In this scenario, the sedating agents should be held for several half-lives prior to paralysis (to avoid a situation where the sedation wears off while the patient is paralyzed).
  • In some situations, paralysis may be desirable while the patient is being sedated (e.g., monitoring for burst suppression in the treatment of status epilepticus, while on anesthetic infusions). In this scenario, the anesthetic infusions should be maintained at a stable dose before and during paralysis.
• 💡 Paralysis without sedation is safe and acceptable – but only if the patient has been carefully assessed and deemed to be deeply and stably comatose.(28187794)

nuts and bolts

• Typically, a single dose of vecuronium (10 mg IV) is administered. This may lead to a period of paralysis of roughly ~30-60 minutes.
• If a shorter duration of paralysis is desired (closer to ~30 minutes), a smaller dose of vecuronium may be utilized (e.g., 0.1 mg/kg).
• Rocuronium may also be utilized at a dose of 0.5-1 mg/kg (depending on the desired duration of paralysis). The choice of vecuronium vs. rocuronium depends on availability and cost – either is equally effective for this use (rocuronium is preferred for intubation when rapid onset is important, but rapid onset is irrelevant here).

---

EEG in encephalopathy or coma – overview

(back to contents)

---

clinical utility of EEG in patients with altered consciousness includes:

• Monitoring depth of sedation (thereby avoiding excess sedation).
• Differentiating delirium from psychiatric disease (e.g., identification of psychogenic unresponsiveness).
• Differentiating delirium from catatonia. 📖

typical findings seen with progressively severe encephalopathy (LaRoche 2018; 28187794)

• (#1) Initial changes may vary, depending on the etiology of encephalopathy:
  • There may initially be desynchronization or the appearance of “fast activity,” with subsequent increases in rhythmicity and voltage (especially delta activity).
  • There may initially be progressive slowing (first in the theta range, and later in the delta range).(28187794)
  • Some etiologies may lead to the emergence of ictal-interictal patterns (e.g., generalized periodic discharges, including triphasic waves 📖).
• (#2) Burst suppression, with duration of suppression becoming longer with deeper levels of encephalopathy.
• (#3) Complete suppression.

EEG patterns that may associate with coma

• Numerous EEG patterns may be seen in comatose patients (e.g., nonconvulsive status epilepticus and various ictal-interictal patterns). The following sections discuss some specific EEG findings which may be seen in coma, but these are certainly not the only findings which may be observed in comatose patients.
• ⚠️ The sections below provide some generalizations about the interpretation of EEG patterns seen in comatose patients.(31307621, 16751721, Sutter 2012) These are not based on high-level evidence, so they are intended primarily to help point clinicians in the right direction – rather than definitively determining the etiology of coma. As always, EEG must be placed within the context of clinical history, neurologic examination, and neuroimaging.

---

beta coma

(back to contents)

---

general comments on accentuated beta activity (“increased fast activity”)

• Normally the resting posterior dominant rhythm is in the alpha range (8-13 Hz).
• Causes of increased beta activity include the following:(31307621)
  • Medications, including light sedation (e.g., benzodiazepine, barbiturate, sympathomimetics, tricyclic antidepressants).
  • Alcohol withdrawal.
  • Hyperthyroidism.
• Among comatose patients with predominant beta activity, roughly two forms of “beta coma” are discernable:

beta coma, favorable type (more common)

• EEG features:
  • EEG may show intermingled alpha activity along with beta activity.
  • EEG may be reactive.
• Causes: intoxication/withdrawal (barbiturates or benzodiazepines), severe hyperthyroidism.

beta coma, unfavorable type (less common)

• EEG features: (Sutter 2012)
  • May show intermingled delta activity along with beta activity.
  • EEG is usually unreactive.
• Causes include: brainstem lesions.

---

alpha coma

(back to contents)

---

The normal awake background EEG pattern is usually composed of a posterior predominant, 8-12 Hz waveform known as alpha waves.  When this pattern instead becomes frontally predominant, and becomes monotonous and unreactive, it is known as an alpha coma pattern. Alpha coma may be divided into roughly three entities:

alpha coma due to intoxication

• Causes include barbiturates, benzodiazepines, anesthetic agents, tricyclic antidepressants.
• EEG features:
  • Alpha activity is often frontally predominant.
  • There is usually some EEG reactivity.
  • There may be superimposed beta activity.(16751721)
• Prognosis is favorable. Patients may transition to a theta pattern prior to normalization.(16751721)

alpha coma due to anoxic brain injury

• EEG often shows:(LaRoche 2018)
  • Diffuse alpha activity that may be maximal over the bifrontal or posterior regions.(31307621)
  • Low amplitude (10-50 uV).
  • Monotonous, without variability.
  • Usually unreactive.
• An alpha coma pattern is strongly linked to a poor outcome.(32562686) This may reflect brainstem injury.(Alkhachroum 2022)
  • Alpha coma may be a transitional state that occurs after burst suppression and prior to electrocerebral silence.(31307621)
  • Patients may also transition between an alpha coma and theta coma (sometimes called an alpha-theta coma).

alpha coma due to brainstem lesions

• EEG may look essentially normal:
  • Normal alpha rhythm with posterior predominance.
  • There can be reactivity.
• Associated with lesions at or just caudal to the pontomesencephalic junction.
• Differential diagnostic considerations for comatose patients with a normal EEG pattern may include locked-in syndrome, catatonia, or psychogenic unresponsiveness.

---

theta coma

(back to contents)

---

more benign theta coma

• Theta-dominant rhythms may result from diffuse cortical dysfunction (e.g., dementia, mild-to-moderate encephalopathy, or severe systemic infection).(Sutter 2012)

theta coma due to anoxic brain injury

• Anoxic brain injury may cause diffuse and unreactive theta activity (often with anterior predominance).
• This usually carries a poor prognosis. Like alpha coma following anoxic brain injury, theta coma may be a transitional state that occurs after burst suppression and prior to electrocerebral silence.(31307621)

---

delta coma

(back to contents)

---

high-voltage delta coma (aka polymorphic delta)

• EEG features:
  • This is defined as a slow background (1-3 Hz delta waves) with high amplitudes (sometimes reaching several hundred microvolts).
  • Delta waves may attenuate with stimulation.(16751721)
  • GPDs with triphasic morphology can occur in some patients.
• Clinical associations of high-voltage delta coma include:
  • Subcortical white matter lesions (focal or unilateral lesions may cause anterior or focal predominance of the delta waves).
  • Severe metabolic encephalopathy.
  • Severe encephalitis.
  • Markedly elevated intracranial pressure.

low-voltage delta coma, favorable type

• EEG features:(Sutter 2012)
  • Delta waves have low amplitude (<20 microvolts).
  • Theta and delta activity with intrusions of alpha and beta activity.(16751721)
  • Reactivity is generally present.
• This may be seen in healthy individuals, or in those with less severe traumatic brain injury.

low-voltage delta coma, unfavorable type

• EEG features:(Sutter 2012)
  • Delta waves have low amplitude (<20 microvolts).
  • Theta and delta activity occurs without intrusions of higher frequency activity.
  • EEG is generally unreactive.
• This may occur in anoxic brain injury or in severe traumatic brain injury.(16751721)

---

spindle coma

(back to contents)

---

EEG findings in spindle coma resemble stages 2-3 non-REM sleep

• Predominant theta and delta background activity occurs with superimposed, frequent, paroxysmal spindle-shaped bursts of 10-14 Hz activity. These spindles are usually bilateral, symmetric, and synchronous.(31307621)
• Intermittent elements of sleep (e.g., K-complexes) may be triggered by stimulation.

causes of spindle coma

• Most closely associated with midbrain lesions (e.g., infarction).(31307621)
• Traumatic brain injury.
• Postictal states.
• Toxic/metabolic encephalopathy.
• Encephalitis and other diffuse cerebral insults.

In the absence of structural brain injury, prognosis is often favorable.

---

epileptiform discharges

(back to contents)

---

Sporadic epileptiform discharges may be seen in a baseline (interictal) EEG in patients with seizures. Epileptiform discharges include the following types:(33475321)

• Spike: clear deviation from background activity with a pointed peak and a duration of 20-70 ms.
• Sharp wave: similar to a spike, but with a duration of 70-200 ms. (The clinical significance of a sharp wave is the same as a spike, so the precise differentiation is unimportant.)
• Polyspike: two or more spikes occurring in a row and lasting <0.5 seconds (if prolonged >0.5 seconds, this might qualify as BIRDs 📖).

Epileptiform discharges should satisfy at least four of the following six criteria.(30214992) These criteria may be useful to emphasize key attributes of epileptiform discharges, while distinguishing them from artefacts:

• Biphasic or triphasic waves with sharp or spiky morphology.
• Asymmetry of the waveform: a sharply rising ascending phase and a more slowly decaying descending phase, or vice versa.
• Spike-wave: A sporadic epileptiform discharge followed by an associated slow after-wave.
• Different wave-duration from the ongoing background activity (either shorter or longer).
  • The epileptiform discharge should be discrete, not just an accentuation of one wave in a sinusoidal wave series.(Spencer 2022)
• Disrupted background activity following the epileptiform discharge.
• Dipole: Distribution of the negative and positive potentials on the scalp suggests a source of the signal within the brain.
  • 💡 The presence of a dipole doesn't necessarily indicate that a spike is epileptiform. Rather this is merely useful to differentiate true brain activity versus artefact.

---

seizures

(back to contents)

---

EEG features of a seizure:

• Series of epileptiform discharges occurring with a frequency of >2.5 Hz.
• Lasts >10 seconds.
• Definite evolution (e.g., postdischarge slowing and/or attenuation).
• Clear onset and offset.

---

BIRDs

(back to contents)

---

brief potentially ictal rhythmic discharges (BIRDs)

• 🐥 These have the features of a seizure, but they are <10 seconds. Specifically:
  • Epileptiform discharges occurring with a frequency >2.5 Hz.
  • Evolution (although this may be difficult to discern in brief abnormalities).
  • Lasting <10 seconds.
• BIRDs are usually lateralized, but may be generalized. In patients with a single focus of acute brain injury, BIRDs are usually localized to that focus.(25678868)
• 🐥 BIRDs may be thought of as “mini-seizures” (patients often also have unequivocal seizures with the same distribution and morphology – simply lasting a bit longer). The 10-second cutoff between BIRDs and unequivocal seizures is an arbitrary cutoff, so BIRDs with a length close to 10 seconds may be nearly indistinguishable from an unequivocal seizure.

causes

• 🐥 BIRDs are one of the most rare patterns, occurring only in 2% of critically ill patients.
• 🐥 Most BIRDs are caused by an acute cerebral injury or chronic seizure disorder.(LaRoche 2018) Among ICU patients with BIRDs, most have focal pathology which is visible on neuroimaging (e.g., stroke or malignancy).(25678868)

clinical implications

• 🐥 BIRDs are associated with a very high risk of seizures (up to 75%), which are mostly nonconvulsive.(32222672) BIRDs may serve as a clinically actionable marker of seizures:(25678868)
  • BIRDs nearly always precede the onset of unequivocal electrographic seizures.
  • In patients with seizures that are controlled, BIRDs are likewise eliminated.
  • Among patients with unequivocal seizures and BIRDs, these usually have the same morphology (implying that they're truly on a continuum of the same underlying phenomena).
• 🐥 Antiseizure medication is generally warranted, along with ongoing monitoring.(30921017)

---

ictal-interictal continuum

(back to contents)

---

definition of the ictal-interictal continuum (IIC)

• This may be roughly regarded as seizure-ish EEG activity that doesn't qualify as an unequivocal seizure, yet there is a reasonable chance that it might be contributing to impaired alertness, other symptoms, and/or neuronal injury.(33475321)
• The following patterns are on the ictal-interictal continuum:
  • Any periodic discharge or spike-wave pattern averaging 1-2.5 Hz over 10 seconds.
  • Any periodic discharge or spike-wave pattern averaging 0.5-1 Hz over 10 seconds with a plus modifier or fluctuation.
  • Lateralized rhythmic delta activity (LRDA) averaging >1 Hz for 10 seconds with a plus modifier or fluctuation.
  • (Note that generalized rhythmic delta activity (GRDA) does not lie on the ictal-interictal continuum).
• 💡 If periodic discharges are associated with a time-locked motor correlate, that alone constitutes nonconvulsive status epilepticus. This would be considered frank seizure (“ictal”) activity, rather than lying within the ictal-interictal continuum.
  • More on the definition of nonconvulsive status epilepticus here: 📖

clinical conundrum of the ictal-interictal continuum

• The ictal-interictal continuum is a purgatory of EEG rhythms which lies between normal rhythms and seizure. The optimal management of these EEG findings is often unclear.
  • Undertreatment could lead to neuronal damage or leave patients in a persistently altered mental status (untreated nonconvulsive status epilepticus).
  • Overtreatment could cause patients to receive unnecessary antiepileptic and anesthetic agents – which could prolong or even precipitate intubation.
• It's often unclear whether EEG patterns are causing harm, or merely reflective of underlying brain dysfunction. However, periodic discharges seem to cause similar metabolic consequences compared to seizure, especially at higher frequencies (>2Hz).(32222672)
• More on the clinical approach here.

---

GPDs – Generalized Periodic Discharges

(back to contents)

---

basics

• GPDs are abnormally synchronized, bihemispheric cerebral activity. GPDs are invariably associated with brain dysfunction (e.g., toxic/metabolic encephalopathy). Most patients are clinically stuporous or comatose.(30214982)
• GPDs are associated with seizures (both focal and generalized), as well as nonconvulsive status epilepticus (NCSE). GPDs associated with plus features (superimposed fast or rhythmic delta activity) or higher frequency (>1.5 Hz) are more strongly associated with seizures (figure below). ~20% of patients may have nonconvulsive status epilepticus.(32222672)

GPDs are usually caused by diffuse neurological insults:(32222672; 29979285, LaRoche 2018)

• Systemic metabolic illness:
  • Hepatic encephalopathy, hyperammonemia.
  • Uremia.
  • Hyponatremia / hypernatremia.
  • Hypoglycemia.
  • Hypothyroidism.
  • Anoxic brain injury: When associated with myoclonus, GPDs may represent an electrographic correlate of myoclonic status epilepticus.(30921017) GPDs on the context of a suppressed background are a very poor prognostic sign. 📖
• Metabolic intoxication or withdrawal:
  • Withdrawal from benzodiazepines, barbiturates, or propofol.(LaRoche 2018)
  • Baclofen.
  • Lithium.
  • Naproxen.
  • Cefepime and other cephalosporins.
  • Valproic acid intoxication.
  • Ketamine or phencyclidine (PCP).
• Infectious/inflammatory etiologies:
  • Sepsis (~1/3 of patients have GPDs).(LaRoche 2018)
  • HSV encephalitis.
  • Steroid-responsive encephalopathy with antithyroid antibodies (SREAT).
  • Hashimoto encephalopathy. 📖
• Other causes:
  • Acute ischemic stroke (AIS).
  • Posterior reversible encephalopathy syndrome (PRES).
  • Traumatic brain injury (TBI).
  • Terminal phase of status epilepticus.
  • GRAW (GPDs related to anesthetic withdrawal). These arise after weaning a patient with status epilepticus off antiepileptic infusions. They are not thought to be an ictal rhythm, even at frequencies up to 4 Hz.(LaRoche 2018)
  • Neurodegenerative disorders (Creutzfeldt-Jakob disease classically causes 1-Hz GPDs; Alzheimer's disease).

additional notes on EEG findings

• One subtype of GPDs is triphasic waves (discussed further at the bottom of this section). Overall, triphasic waves have clinical implications similar to those of other types of GPDs.
• The background is usually suppressed or slowed (in the delta-theta range).
• GPDs are not the same as bursts (GPDs may have up to three phases, whereas bursts have >3 phases and last longer).
• Patients may manifest with multiple different types of periodic activity (e.g., GPDs combined intermittently with lateralized periodic discharges (LPDs) or bilateral independent periodic discharges (BIPDs)).(LaRoche 2018)

clinical approach to the patient with GPDs

• Evaluation of the cause: This may include a review of medications, as well as laboratory studies, to evaluate for metabolic derangements (consider the causes of GPDs listed above). Treatment of any specific cause may be the most important intervention.
• Monitoring for electrographic seizures: Prolonged EEG monitoring for seizures is generally advisable.(Wijdicks 2019)
• Antiepileptic therapy?
  • It's often unclear whether GPDs require specific treatment.(29979285)
  • Triphasic waves may sometimes show electrocardiographic and clinical improvement following antiepileptic therapy, implying an ictal etiology of the patient's altered mental status.(29650639) One multi-institutional retrospective series of 64 patients with triphasic waves reported a higher response rate (42% vs. 19%) when treated with a non-benzodiazepine antiepileptic (mostly levetiracetam).(26013921)
  • When in doubt, a trial of antiepileptic therapy may be considered (more on this here 📖).

GPDs with triphasic morphology (triphasic waves)

• Triphasic waves are a specific subtype of generalized periodic discharge with the following characteristics:
  • Three phases (negative – positive – negative), each of longer duration than the preceding wave.
  • Frontal predominance.
• Triphasic waves have traditionally been felt to be an epiphenomenon of metabolic encephalopathy, with little additional importance. However, some recent research has raised the question of whether triphasic waves may be contributing to metabolic encephalopathy – and thus merit specific therapy.
• Triphasic waves should be managed similarly compared to other GPDs, for the following reasons:
  • (1) Inter-rater variability in identifying triphasic waves is often limited.(LaRoche 2018)
  • (2) The clinical implications of triphasic waves seems to be similar to other GPDs. For example, triphasic waves carry the same risk of seizures compared to other GPDs.(29262437; LaRoche 2018)

---

LPDs – Lateralized Periodic Discharges

(back to contents)

---

basics

• LPDs are usually caused by acute/subacute, ipsilateral structural brain lesions.(29262437) LPDs usually resolve after 8-10 days, rarely persisting beyond a few weeks.(30214982) However, LPDs may occasionally occur due to chronic pathology (e.g., as a postictal feature, or acute reactivation of an underlying chronic structural lesion).(LaRoche 2018; 35393960)
• Among all ictal-interictal patterns, LPDs are the most closely associated with seizure (with ~65% of patients having seizures during their acute illness, often nonconvulsive).(32222672, 29666958) Risk of seizure is higher in the presence of additional features (superimposed fast or rhythmic delta activity) or higher frequency LPDs (>1.5-2 Hz). LPDs are also associated with a risk of developing epilepsy.(LaRoche 2018)
• Aside from their association with unequivocal seizures, LPDs themselves may represent a form of focal nonconvulsive status epilepticus (NCSE). Some evidence to support this:
  • (1) LPDs occurring near the motor cortex are associated with time-locked motor activity (epilepsia partialis continua, discussed further here: 📖).(29979285) LPDs in other areas probably cause symptoms that are less obvious (e.g., eye deviation, aphasia, hemianopsia, or sensory disturbances). In patients with LPDs and altered consciousness, it may be difficult to discern whether the LPDs are causing impaired mental status.(LaRoche 2018)
  • (2) Some studies have found that LPDs correlated with glucose hypermetabolism seen on PET scans. LPDs >2 Hz have been associated with focal tissue hypoxemia.(35393970) This may imply that LPDs could cause neuronal injury.(LaRoche 2018)

causes of LPDs (usually focal pathology) (32222672)

• Stroke
  • Ischemic stroke (the most common cause of LPDs overall).
  • Intracerebral hemorrhage.
  • Subarachnoid hemorrhage (in this context, LPDs predict the development of delayed cerebral edema).(29262437)
• CNS infection (e.g., abscess, HSV encephalitis).
  • Temporal LPDs are a hallmark finding of HSV encephalitis.(Wijdicks 2019)
• Brain tumor.
• Autoimmune encephalitis.(LaRoche 2018)
• Traumatic brain injury (TBI).
• PRES may cause LPDs, typically in a posterior distribution.(34798964)
• Epilepsy.
• Less common causes:(LaRoche 2018)
  • Postanoxic and toxic-metabolic encephalopathies.(29262437)
  • Creutzfeldt-Jakob disease (CJD).
  • Demyelinating diseases.
  • Migraine.
  • Mitochondrial encephalopathy with lactic acidosis and stroke-like episodes (MELAS).
  • Advanced stages of progressive myoclonic epilepsy.
  • Neurosyphilis.
  • SESA syndrome (subacute encephalopathy with seizures in alcoholics). 📖

some EEG characteristics

• LPDs vary in precise location (e.g., they may be maximal in a specific location).
• LPDs usually associate with EEG evidence of ipsilateral cerebral dysfunction (focal slowing, loss of posterior dominant rhythm).(LaRoche 2018)
  • LPDs alternating with EEG suppression are more likely to represent an ictal phenomenon, which may be very difficult to treat (e.g., focal nonconvulsive status epilepticus).(34034086)

clinical approach to the patient with LPDs

• Evaluation of the cause:
  • LPDs are usually caused by focal pathology, so neuroimaging should generally be performed.
  • Depending on the clinical context, additional causes may need to be considered and investigated (e.g., HSV encephalitis).
  • The most important intervention is to determine the underlying cause of LPDs and treat it (if possible).
• Monitoring for electrographic seizures: Given the strong association between LPDs and seizures, prolonged EEG monitoring for seizures is generally advisable.
• Seizure prophylaxis: Given that ~65% of patients with LPDs have seizures, most patients with LPDs should receive seizure prophylaxis (at least in the acute phase of their illness).(LaRoche 2018)
• Elimination of LPDs:
  • In patients with a time-locked motor correlate, LPDs represent electroclinical nonconvulsive status epilepticus (NCSE). In this situation, attempts are usually made to suppress LPDs (further discussion in the section on epilepsia partialis continua: 📖).
  • In patients without an overt clinical correlate, management is unclear. If there is concern that LPDs are causing symptoms such as impaired consciousness, a time-limited trial of antiepileptic therapy may be warranted (discussed further here: 📖). Clinical context is important (e.g., if the patient is improving, then more conservative management may be best). The use of propofol or midazolam infusions to suppress LPDs is generally inadvisable.

---

LRDA – Lateralized Rhythmic Delta Activity

(back to contents)

---

clinical significance

• LRDA is seen in 5% of critically ill patients.(32222672)
• LRDA overall seems to have a similar clinical implications as LPDs 📖. Some evidence from intracranial recordings suggests that LRDA on scalp EEG may actually reflect intracranial periodic discharges – which could explain the clinical similarities of these two patterns.(LaRoche 2018)
  • LRDA commonly coexists with LPDs (a combination that carries a very high risk of acute seizure).(LaRoche 2018)
• LRDA is usually due to an underlying focal lesion (although occasionally the lesion may occur on the contralateral side of the brain).(29666958)
• Most patients have moderate alteration of consciousness and focal neurological deficits.(LaRoche 2018)
• LRDA is associated with a risk of seizure in ~50% of patients, especially nonconvulsive status epilepticus.(29666958) Seizure risk seems to emerge at frequencies >1.5 Hz (figure below).

causes of LRDA & clinical approach

• Causes are generally similar to the causes of LPDs.
• Management is generally similar to patients with LPDs (discussed in the section above: 📖).

---

BIPDs – Bilateral Independent Periodic Discharges

(back to contents)

---

clinical significance

• BIPDs are less common than LPDs. Compared to LPDs, BIPDs are overall associated with a worse prognosis (including a higher mortality).(29262437)
• BIPDs may be less closely associated with seizure as compared to LPDs, but they may be more frequently associated with coma.

causes of BIPDs

• BIPDs most commonly occur in acute/subacute structural brain injury (similar to LPDs). Unlike LPDs, they are more often seen in the setting of bilateral brain lesions.(29666958)
• Overall, BIPDs and LPDs are caused by similar disease processes. However, BIPDs may be more likely to occur with diffuse processes (e.g., metabolic abnormalities or anoxic brain injury).
• Causes of BIPDs include:(29262437)
  • Stroke.
  • CNS infection (e.g., herpes simplex encephalitis).
  • Tumor.
  • Posterior reversible encephalopathy syndrome (PRES).
  • Metabolic disturbances (e.g., severe hypoglycemia).(33456874)
  • Anoxic brain injury (in this context, BIPDs are an extremely poor prognostic sign).(LaRoche 2018)

clinical approach to the patient with BIPDs

• BIPDs are less common than LPDs, so there is less written about this specifically.
• A management approach similar to LPDs may be reasonable (discussed above 📖).

---

SIRPIDs – Stimulus Induced Rhythmic, Periodic, or Ictal Discharges

(back to contents)

---

causes of SIRPIDs

• SIRPIDs are most often seen in acute brain injury. SIRPIDs may be seen in up to 20% of patients in a neurocritical care unit.(Spencer 2022)
• SIRPIDs are reported in wide range of conditions:
  • Stroke, intracranial hemorrhage, subarachnoid hemorrhage.
  • Traumatic brain injury.
  • Hypoxic-ischemic encephalopathy.
  • Toxic/metabolic disorders (e.g., drug toxicity, hyponatremia).
  • Neurodegenerative disorders (e.g., Creutzfeldt-Jakob disease).(29262437)

clinical significance

• SIRPIDs appear a bit more benign than spontaneous patterns:
  • SIRPIDs have a somewhat lower association with seizure, as compared with spontaneous ictal-interictal continuum patterns.(32222672)
  • SIRPIDs are usually not associated with any clinical correlate, although some reports have described focal motor activity.(29666958)
• 🔑 Most experts recommend treating SIRPIDs in the same manner as spontaneous patterns (as discussed in more detail above – with different approaches for different patterns).(LaRoche 2018)

---

GRDA – Generalized Rhythmic Delta Activity

(back to contents)

---

basics of GRDA (the rhythm formerly known as FIRDA)

• GRDA is one of the most commonly encountered patterns in acutely ill patients.
• Clinically GRDA is usually correlated with some degree of encephalopathy, usually mild.(LaRoche 2018) It's associated with a relatively favorable prognosis.
• GRDA has not been associated with increased seizure risk (regardless of frequency or plus modifiers).(32222672) Therefore, GRDA does not lie on the ictal-interictal continuum.(29666958)

causes of GRDA

• Encephalopathy (e.g., due to traumatic, toxic-metabolic, or inflammatory disorders).
  • Anti-NMDA encephalitis (may cause prolonged runs of GRDA).
  • Especially frontally predominant GRDA is considered a nonspecific pattern reflective of encephalopathy.(29666958)
• Prior seizure (with GRDA in the postictal period), or generalized epilepsy.
  • Temporally predominant GRDA (previously termed TIRDA) in particular is associated with temporal lobe epilepsy. Clinically, this has a significance similar to temporal epileptiform discharges.(Spencer 2022)
  • (Occipitally predominant GRDA is associated with absence seizures, but this is usually not seen in adults.)
• Elevated pressure within the third ventricle, cerebral edema.
• Structural brain lesions (e.g., subcortical lesions, tumors).
• Neurodegenerative disorders (e.g., corticobasal degeneration, progressive supranuclear palsy, Creutzfeldt-Jakob disease).(31307621)

---

extreme delta brush

(back to contents)

---

extreme delta brush requires two basic components 🖌

1. Rhythmic delta activity -or- periodic discharges that have a blunt delta-wave morphology.
2. Superimposed fast activity (+F) that has a stereotyped relationship to the delta waves.

clinical significance:

• Extreme delta brush is highly suggestive of anti-NMDA receptor encephalitis.📖 It is associated with prolonged disease course and status epilepticus.

---

CAPE – Cyclic Alternating Pattern of Encephalopathy

(back to contents)

---

• CAPE is defined as spontaneously alternating between two background patterns in a regular manner, as shown above. Background patterns may include rhythmic and periodic patterns.(33475321)
• CAPE may be seen in comatose patients, among whom this is associated with a favorable prognosis.(18090521)

---

podcast

(back to contents)

---

Follow us on iTunes

---

questions & discussion

(back to contents)

---

To keep this page small and fast, questions & discussion about this post can be found on another page here.

• Under-interpretation or over-interpretation of the EEG.
• The art of neurocritical care involves determining how aggressive to be in treating various EEG abnormalities. In the absence of any high-quality evidence, this often requires personalization, integration with clinical information, and some trial and error. Attempting to suppress all scary EEG findings may cause severe iatrogenic harm (e.g., due to prolonging the duration of anesthetic infusions).

Guide to emoji hyperlinks

• = Link to online calculator.
• = Link to Medscape monograph about a drug.
• = Link to IBCC section about a drug.
• = Link to IBCC section covering that topic.
• = Link to FOAMed site with related information.
• 📄 = Link to open-access journal article.
• = Link to supplemental media.