Quantitative EEG (qEEG)

1. Bedside vignettes: why this matters in the PICU#

Vignette A. SAH day 6, ADR falling#

A 14-year-old is on day 6 after a ruptured AVM with secondary SAH. The clinical exam is unchanged, the daily TCD is borderline (MFV 140, Lindegaard 2.8), and ICP is well controlled. The cEEG / qEEG technologist updates the DSA: over the last 6 hours, the right hemisphere ADR has fallen from 0.42 to 0.28, a 33% drop. Total power on the right is stable; the asymmetry index has crossed the 0.3 alert threshold. There is no overt clinical sign. The team escalates: triple-H haemodynamics, repeat CT angio, and decision-prep for endovascular treatment. The qEEG caught the ischaemic signature 12 hours before the next angio confirmed early vasospasm progression. 1 2

Vignette B. Neonatal seizure burden on cEEG / qEEG#

A 5-day-old term infant with HIE on day 3 of normothermic rewarming is on cEEG. The bedside DSA shows three discrete bright vertical "fingers" (each 2 to 4 minutes long) over the last hour, occupying about 20% of the trace. The amplitude-integrated EEG (aEEG) trace shows abrupt elevations in the upper margin during the same intervals. These are electrographic seizures, recurrent and meeting the ACNS quantitative definition of neonatal status epilepticus (> 50% of any 1 h epoch). Treatment is escalated from phenobarbital to phenytoin then to midazolam infusion with bedside DSA / aEEG endpoints. 3 4

Vignette C. Sedated TBI with paradoxically high ADR#

A 12-year-old with severe TBI is on continuous midazolam and fentanyl, ICP 18 mmHg, BIS 35. The cEEG is markedly suppressed with intermittent bursts. The bedside DSA shows a fairly bright "alpha band" (8 to 13 Hz) at all leads, and the ADR is 0.6 (high). The neurophysiologist explains: this is not preserved alpha-rhythm-of-arousal, but a sedative-induced anterior alpha pattern (propofol- and benzodiazepine-induced frontal alpha), and the ADR here is uninformative about cortical metabolism. qEEG values must always be interpreted with knowledge of sedation, age, and the raw EEG context. 5 6

2. What qEEG is, and what it is not#

Continuous EEG (cEEG) generates roughly 24 channels × 256 Hz × 86,400 s = ~500 million samples per day. A neurophysiologist cannot scan this volume of raw EEG continuously, and a bedside nurse or intensivist usually cannot read raw EEG at all. qEEG is the family of data-reduction algorithms that turn this firehose into a manageable, trendable bedside display.

The key qEEG outputs.

• DSA (density spectral array): a colour-coded time-frequency map per channel. Time on the x-axis (typically the last 6 to 24 h), frequency on the y-axis (0 to 30 Hz), power encoded as colour intensity. The single most useful qEEG display.
• ADR (alpha-delta ratio): instantaneous ratio of 8 to 13 Hz power to 1 to 4 Hz power. Falls with cortical ischaemia (alpha disappears, delta predominates). The most validated qEEG metric for DCI detection in SAH.
• BSR (burst suppression ratio) / PSI (power suppression index): proportion of time spent in low-amplitude suppression. Used for sedation titration (BSR 30 to 60% is a common target in pentobarbital coma) and post-arrest prognostication.
• Total power: amplitude in microvolts across the whole spectrum. Falls during global ischaemia or deep sedation.
• Asymmetry index: difference in spectral power between the hemispheres; rises with unilateral lesion or ischaemia.
• SEF95: frequency below which 95% of power lies; falls in slowed EEG (sedation, encephalopathy, ischaemia).

Three things follow.

qEEG does not replace raw EEG review. Artefact (chewing, sweat, electrode pop, infusion-pump 60 Hz noise) can dominate qEEG metrics. A neurophysiologist must validate the raw trace at least daily and flag artefacts. qEEG metrics in isolation can mislead. 6 7

qEEG amplifies what raw EEG already shows. The DSA reveals trends (slow drift in ADR, asymmetry building over hours, recurrence of seizures in the night) that are invisible to short single-snapshot raw-EEG reads.

Pediatric qEEG requires pediatric-aware interpretation. Normal pediatric EEG is faster (alpha frequency rises through childhood, peaking at ~10 Hz by mid-school age), so the "alpha band" in a 6-month-old is centred lower than in a 12-year-old. Suppression patterns and burst morphology also differ. 8 4

3. Anatomy and montage: the substrate of qEEG#

The qEEG display sits on top of a properly placed multi-channel EEG. Three practical layers:

1. Electrode placement: international 10-20 system (or 10-10 in research / refractory work). Most pediatric ICU recordings use 16 to 21 electrodes. Neonatal montages often use a reduced "neonatal montage" (~10 electrodes) per ACNS 2013 guidelines. 8
2. Reference: a reference electrode is needed for monopolar montages (Cz / linked-ears / average reference). Bipolar montages subtract adjacent pairs and are reference-independent.
3. Computation: the qEEG software computes the desired metrics (DSA, ADR, BSR, SEF, asymmetry) in 2 to 10 s epochs and updates the display in near real time.

Skin preparation is the single biggest determinant of trace quality. Electrode impedance < 5 kΩ at every channel is the goal. Sub-optimal prep gives artefact-laden traces whose qEEG metrics are uninterpretable.

4. The DSA: anatomy of a spectrogram#

How to read a DSA at the bedside.

1. Look at the colour gradient: bright yellow / red = high power, dark blue = low power.
2. Bands of horizontal colour correspond to dominant frequency bands. A bright band at 8 to 13 Hz is alpha, 4 to 7 Hz is theta, 1 to 4 Hz is delta, 13 to 30 Hz is beta.
3. Vertical bright stripes are typically seizures (rhythmic activity across all frequencies for a short time).
4. Slow drift in colour over hours is the most diagnostically interesting signal: a gradual loss of high-frequency colour is encephalopathy; a gradual asymmetry building over hours is unilateral ischaemia; a recovering bright alpha band post-arrest is reassuring.
5. Asymmetry between hemispheric panels is the focal-ischaemia signal.

The four common qEEG metrics on a single screen.

5. The numbers: what to record at the bedside#

The headline display is the DSA. The bedside flow sheet adds the derived numbers (ADR, BSR, asymmetry) at hourly intervals or at clinically defined epochs. The full report (with raw EEG review by neurophysiology) is daily or twice-daily.

6. What is normal? Age-banded reference patterns#

8 6 3.

7. What is abnormal? Pattern library#

Decision tree: "what is the DSA telling me?"#

flowchart TD
  DSA[Bedside DSA] --> Asym{Asymmetry > 0.3?}
  Asym -->|Yes| Focal[Focal process; imaging]
  Asym -->|No| ADR{ADR drop > 25% over 6 h?}
  ADR -->|Yes| Isch[Cortical ischaemia; investigate]
  ADR -->|No| Bgd{Background continuous?}
  Bgd -->|Yes, age-appropriate| Norm[Continue]
  Bgd -->|No, suppressed| Sed{Sedation on?}
  Sed -->|Yes| Titr[Titrate to BSR target]
  Sed -->|No| Brain[Severe injury; multimodal prognostication]

8. Try it: interactive widgets#

9. qEEG-driven management decisions#

9.1 Status epilepticus titration#

In refractory SE, qEEG provides the bedside endpoint for continuous-infusion sedatives (midazolam, pentobarbital, ketamine). Common targets:

• Burst suppression with BSR 30 to 60% for pentobarbital coma.
• Seizure freedom (no electrographic seizures for 24 h) before weaning.
• Background reactivity is monitored throughout.

The ESETT and ECLIPSE-SE trials established the bedside use of cEEG / qEEG endpoints for SE management in pediatric and adult patients. 9 10

9.2 SAH-DCI surveillance#

ADR drops detect cortical ischaemia 12 to 48 h before clinical signs of DCI. Combined with daily TCD and clinical exam, qEEG-driven escalation includes BP optimisation, fluid status review, and angiography. The Claassen 2004 cohort and Sandsmark 2024 multicentre data are the canonical evidence. 1 11 2 12

9.3 Post-cardiac-arrest prognostication#

qEEG features at 24 to 72 h post-arrest are part of the multimodal prognostication framework. Suppressed background, sustained burst suppression with low burst variability, GPDs, and absent reactivity correlate with poor outcome. Continuous near-normal background is reassuring. Single-modality prognostication is forbidden by current pediatric guidelines. 13 14

9.4 Sedation depth in the paralysed patient#

In a paralysed patient on continuous sedation (e.g., refractory ICP, refractory SE), qEEG / BIS-derived metrics (SEF, BSR) provide a sedation-depth proxy when the clinical exam is uninformative. BIS is a single-channel forehead derivative; qEEG with full montage gives more information at the cost of complexity. 5

9.5 Detection of NCSE in unexplained coma#

Up to 20% of comatose ICU patients have NCSE that explains some or all of their depressed exam. cEEG with qEEG bedside trends catches this and supports immediate treatment. The ACNS guidance recommends cEEG (often 24 to 48 h) for any unexplained coma. 6 7

10. Clinical contexts: qEEG across acute brain injuries#

10.1 Aneurysmal SAH and DCI#

The most validated indication. ADR drop > 25% over 6 hours, sustained, in the DCI window (days 3 to 14) is sensitive and specific for evolving cortical ischaemia. Claassen 2004 enrolled 34 patients; sensitivity for DCI ~70%, specificity ~80%; subsequent multicentre validation (Sandsmark 2024) confirmed in 200+ patients. 1 11 2 15 Pair with daily TCD and clinical exam.

10.2 Severe TBI#

cEEG / qEEG is part of the BTF / pBTF management bundle, primarily for detection of NCSE (which occurs in up to 25% of severe TBI patients) and seizure burden quantification. qEEG-derived BSR is a bedside endpoint when burst suppression is chosen for refractory ICP. 16 17 18

10.3 Pediatric AIS#

In hyperacute AIS, cEEG / qEEG detects post-recanalisation seizures (which occur in up to 15% of pediatric AIS, especially with cortical infarcts), monitors hyperperfusion-related cortical irritation, and tracks recovery of normal background over days to weeks. 19 20

10.4 HIE and post-cardiac arrest#

In neonatal HIE, aEEG / qEEG is the standard bedside surveillance tool: it tracks background evolution from suppression / burst-suppression / discontinuous / continuous as the infant rewarms and either recovers or progresses. Sleep-wake cycling at 24 to 48 h is a strong positive prognostic sign. 3 4 In pediatric post-arrest, qEEG at 24, 48, 72 h supports the multimodal prognostication. 13 14

10.5 Pediatric ECMO#

Subclinical seizures occur in 5 to 20% of pediatric ECMO patients. ELSO neurological guidelines recommend cEEG for at least 24 h on initiation of ECMO, with qEEG / aEEG bedside trends for ongoing surveillance. 21 22

10.6 Meningitis and encephalitis#

cEEG / qEEG identifies NCSE, focal cortical irritation (LPDs in herpes encephalitis classically over the temporal lobes), and the slowly recovering background of meningoencephalitis. ACNS guidelines recommend cEEG in any encephalitis with depressed sensorium. 23 24 6

10.7 Brain-death determination#

EEG is an ancillary test in some jurisdictions for brain-death determination. The required pattern is electrocerebral silence: amplitude < 2 µV across all channels for 30 minutes, in the absence of confounders. qEEG (total power suppression) provides a quantitative complement. The pediatric brain-death criteria still require the standard clinical exam and apnoea test. 25 26

10.8 DKA cerebral oedema#

Less validated; cEEG is not routine in DKA. In severe DKA with depressed mental status, cEEG can detect NCSE as a contributor to depressed consciousness, which is treatable. The exam, head CT, and clinical course remain primary in DKA cerebral oedema management. 27 28

10.9 Refractory status epilepticus#

The defining indication. Bedside qEEG / DSA, paired with raw cEEG review, drives titration of continuous infusion (midazolam, pentobarbital, ketamine) to seizure freedom or to a defined BSR. The ESETT trial (first-line second-line agents) and the ECLIPSE-SE trial (third-line refractory SE) are the canonical pediatric evidence. 9 10 29

11. Multimodal integration: qEEG in the MMM/MNM stack#

12. Setup and technique#

12.1 Equipment#

• Multi-channel EEG system with at least 8 to 21 electrodes (more for pediatric, fewer acceptable for emergent neonatal).
• Skin preparation kit: alcohol, abrasive paste, conductive gel, electrodes.
• Long-term electrodes: cup electrodes attached with collodion or paste for sessions > 24 h.
• qEEG software: vendor-specific or open-source (Persyst, ICMplus, BrainVision, EDFbrowser).
• Bedside display: ideally beside the patient, visible to nursing.

12.2 The recording#

1. Skin prep: clean each electrode site with alcohol; abrade gently with paste; apply conductive gel.
2. Electrodes: place per 10-20 system; double-check left vs right, anterior vs posterior; check impedances < 5 kΩ.
3. Reference and ground: typically Cz or linked-ears or average; ground typically at Fpz.
4. Sampling rate: typically 256 to 512 Hz.
5. Filters: low-frequency 1 Hz, high-frequency 70 Hz, notch 50 or 60 Hz.
6. Calibration / biocalibration: confirm signal entry, polarity, eye-blink artefact recognisable.
7. Begin recording; the qEEG software computes DSA / ADR / BSR in near real time.

12.3 Quality control#

• Impedance check at start and at any sudden change in trace quality.
• Artefact identification: chewing, sweating, electrode pop, infusion-pump 60 Hz noise, ventilator artefact. The neurophysiologist annotates these so qEEG metrics are not derived from artefact.
• Daily neurophysiology review: raw trace review by a neurophysiologist at least daily, with reports back to the ICU team.
• Re-prep every 24 to 48 hours, more often in sweaty / paediatric / restless patients.

12.4 Pediatric-specific tips#

• Smaller electrodes for neonates and infants; use neonatal montage (~10 electrodes per ACNS 2013).
• Hair: dense pediatric hair complicates electrode adhesion; use small cup electrodes with conductive paste; some units use disposable sub-dermal needle electrodes in selected neonates.
• Movement artefact: toddlers and unsedated children generate frequent movement artefact; the qEEG team must be conservative about interpreting movement-rich epochs.
• Sleep-wake cycling: an important normal finding in neonates and infants; the qEEG / aEEG bedside display should show diurnal variation.

12.5 The bedside conversation#

The qEEG report to the bedside team should include: the dominant background pattern, the ADR trend over the last 6 to 12 h, the BSR / continuity, the asymmetry index, the seizure burden, the reactivity, and the artefact load. The neurophysiology team's interpretation is the ground truth; the qEEG metrics on screen are signals to look harder.

12.6 When qEEG misleads#

• Sedation effect on alpha: propofol and benzodiazepines induce anterior alpha; the ADR can be misleadingly high.
• Hypothermia: depresses background, lowers ADR, can mimic post-arrest poor prognosis without being one.
• Artefact from ICP wave: pulsatile artefact from the EVD or scalp wound can appear as low-frequency power.
• Movement and EMG: rises with sympathetic surge; appears as high-frequency power, can be confused with seizures.
• Electrode disconnection: zero-amplitude on one channel masquerades as suppression.

13. Pitfalls#

• qEEG without raw-EEG review: the most common error. Always have a neurophysiologist validate the raw trace and annotate artefacts.
• Single-snapshot ADR: trend over 6 to 12 hours is the diagnostic signal, not one number.
• Confounding sedation: midazolam, propofol, and ketamine all alter qEEG metrics; document sedation status with every interpretation.
• Confounding hypothermia: 33 to 34 °C reduces overall power and slows the dominant frequency; do not prognosticate during hypothermia.
• Artefact contamination: ventilator, infusion pumps, chewing, electrode pop, EMG; clean traces are essential.
• Reduced montage in neonates: the trade-off between practicality and information; document the montage with every report.
• Pediatric "normal" is age-dependent: the alpha-frequency band moves through childhood; use age-banded comparisons.
• Comparing absolute values across centres or devices: vendor-specific algorithms (Persyst vs ICMplus vs others) compute differently; compare within a single setup over time.
• Over-reading single subtle asymmetries: small asymmetries are common and often artefactual; require sustained asymmetry over hours.
• Failure to communicate: qEEG / cEEG that runs without being looked at is useless; ensure daily neurophysiology rounds with the ICU team.

14. Combine with…#

• cEEG / EEG: the raw trace that qEEG sits on top of.
• aEEG: the simpler bedside envelope, often paired with full qEEG.
• BIS: a single-channel proprietary sedation-depth proxy; complementary in the paralysed patient.
• TCD: vasospasm + cortical ischaemia (ADR) is the SAH bundle.
• Pupillometry: multimodal post-arrest prognostication.
• Evoked potentials: SSEP + qEEG for post-arrest.
• Microdialysis: cortical biochemistry + cortical electrical activity.

15. Evidence summary#

16. Recent literature (2022 to 2025)#

• Sandsmark 2024: multicentre validation of qEEG ADR drop as DCI early-warning in 200+ SAH patients. 2
• Williams 2024 qEEG review: integration of qEEG into MNM, including ML-augmented interpretation. 12
• Foreman 2022 ICU qEEG review: pragmatic guide to bedside interpretation, artefact recognition, and integration with multimodal monitoring. 5
• Sansevere 2023 pediatric / neonatal cEEG review: framework for cEEG use in PICU and NICU. 4
• Naim 2023 pediatric post-arrest brain injury: qEEG in pediatric multimodal prognostication. 14
• ACNS 2021 updated standardised terminology (Hirsch et al): the lingua franca of qEEG reporting. 7

17. Self-check#