Evoked potentials (SSEP and BAER)

1. Bedside vignettes: why this matters#

Vignette A. Post-arrest 5-year-old at 72 hours#

A 5-year-old after a 20-minute submersion arrest. Day 3, off cooling, off sedation. Pupils equal, sluggish. Bilateral motor extension to pain. The family asks for a prognosis. The neurophysiology team records median-nerve SSEPs: bilateral N20 absent; peripheral N9 and cervical N13 present (technical recording clean). Combined with the neurological exam and a markedly suppressed cEEG, the prognosis is communicated as poor. The decision around WLST is informed, not made, by this single test. 1 2 3

Vignette B. Adolescent brain-death evaluation#

A 16-year-old with a devastating cerebral haemorrhage from a ruptured AVM. Clinical exam is consistent with brain death. The team plans the formal determination. BAER recording: wave I present, waves III–V absent bilaterally, the pattern of brainstem electrophysiologic silence. This is consistent with the clinical findings and supports the determination, but it does not replace the apnoea test and the clinical examination required by the World Brain Death Project framework. 4 5

Vignette C. Neonatal HIE, the BAER threshold story#

A 38-week neonate with severe HIE, day 3 post-rewarming. BAER recorded as part of the neuro-prognostication workup: wave V present but markedly delayed (latency 8.2 ms, expected < 6.5 ms for term newborn); waves I and III preserved. This represents subcortical conduction delay typical of moderate-to-severe HIE; outcome at 18 months tracks with this finding, less catastrophic than wave V loss, more concerning than normal latencies. Pair with the aEEG, MRI, and clinical exam for the full prognostic call. 6 7

2. What evoked potentials are, and what they are not#

Evoked potentials are stimulus-locked, time-averaged electrophysiological signals: the brain's response to a defined sensory stimulus, recorded from scalp or cortical electrodes, with the random EEG noise averaged out by repeating the stimulus hundreds to thousands of times.

The mathematical idea:

where each EEG_i(t) is the recorded scalp signal in the 50–500 ms window following the i-th stimulus, and N is typically 500–2000 stimuli per channel.

The averaging works because the evoked response is time-locked to the stimulus (it occurs at the same latency on each trial) and the EEG background is not: the random EEG signals average toward zero, while the stimulus-locked response sums coherently.

2.1 The SSEP#

• Stimulus: 4-Hz electrical pulse at the median nerve at the wrist (or posterior tibial nerve at the ankle). Current 5–25 mA, sufficient to produce a small thumb twitch.
• Recording sites: peripheral (Erb's point, N9), cervical (over C7, N13), and cortical (over the contralateral parietal scalp, C3' or C4', N20 and N20-P25).
• Pathway tested: peripheral nerve → dorsal columns of the spinal cord → medial lemniscus → ventroposterolateral thalamus → primary somatosensory cortex (S1).
• Bedside read: N20 present (pathway intact, cortical function present); N20 absent in a comatose patient with normal peripheral and cervical recordings is highly specific for catastrophic cortical injury.

2.2 The BAER#

• Stimulus: click (60–80 dB SPL), 11.1-Hz rate, monaural with contralateral masking.
• Recording sites: vertex (Cz) referenced to ipsilateral mastoid.
• Pathway tested: cochlea → cochlear nerve → cochlear nucleus → superior olivary complex → lateral lemniscus → inferior colliculus.
• Five canonical waves (I–V): I from cochlear nerve, II from cochlear nucleus, III from superior olivary, IV from lateral lemniscus, V from inferior colliculus.
• Bedside read: progressive loss of waves III→V is the pattern of progressive brainstem failure; isolated wave I with absent III–V is consistent with brainstem electrophysiologic silence.

What EPs do well#

• Pathway specificity: SSEPs test the dorsal-column-cortex pathway; BAERs test the cochlea-midbrain pathway. Each is highly specific for its pathway.
• Robustness to sedation: SSEPs and BAERs are far less affected by sedation than EEG; you can interpret them in deeply sedated patients.
• Quantitative endpoints: presence/absence of N20 is binary and inter-rater agreement is high.
• Post-arrest prognostication: bilateral absent N20 is the most specific bedside test for catastrophic cortical injury after cardiac arrest. 1 2

What EPs cannot do#

• Test the whole brain: SSEPs test the dorsal-column-cortex pathway only; large frontal or temporal lesions can be present with preserved N20.
• Detect seizures: EPs are stimulus-locked; ictal activity is not.
• Be used in isolation: prognostic calls always combine EP with clinical exam, cEEG, imaging, and (for post-arrest) NPI.
• Be performed reliably with movement / shivering / agitation: technical quality is everything.
• Test motor pathway non-invasively in the ICU: MEPs are an intraoperative tool, not a bedside ICU monitor.

3. Anatomy and electrode placement#

3.1 SSEP electrode placement#

Median-nerve SSEP is the standard for ICU coma prognostication. Posterior tibial nerve SSEPs (recorded at L4 and over the vertex) can be added for testing the long ascending pathway.

3.2 BAER electrode placement#

Stimulus delivered monaurally through an insert earphone, contralateral ear masked with white noise.

3.3 Pediatric considerations#

• Smaller heads require careful electrode positioning; the C3' / C4' landmarks scale with head circumference.
• Higher impedances in neonatal skin; longer skin prep tolerable.
• Stimulus intensity: titrate to a small thumb twitch (median) or visible foot twitch (tibial); lower currents than adult.
• Click intensity for BAER in newborns: 60 dB nHL is appropriate for screening; 80 dB nHL for full brainstem evaluation.

4. The signal: waveform anatomy#

4.1 SSEP waveform#

The median-nerve SSEP, recorded over the contralateral C3' or C4' scalp referenced to Fz, shows a characteristic peak structure:

The N20-P25 complex is the bedside cortical read. The amplitude (typically 1–5 µV) is small but reproducible across two independent recordings of 500–2000 stimuli each.

4.2 BAER waveform#

The auditory click produces five waves in the first 10 ms post-stimulus:

The interpeak latencies (I–III, III–V, I–V) carry the localising information:

• Prolonged I–III: lower brainstem (pontomedullary) lesion
• Prolonged III–V: upper brainstem (pontomesencephalic) lesion
• Prolonged I–V: overall brainstem conduction delay
• Loss of III–V with preserved I: brainstem electrophysiologic silence

5. The numbers to record: the EP six-pack#

Every interpretation pairs the EP read with time since insult, temperature (TTM affects latencies), sedation regimen, and the clinical exam.

6. What is normal? Age-banded reference#

6.1 SSEP latencies#

6.2 BAER latencies#

Sources: 8 1 9. Latencies depend on stimulus intensity, body temperature, and (in BAER) middle-ear status.

7. What is abnormal? Pattern library#

7.1 SSEP patterns#

7.2 BAER patterns#

Decision tree: post-arrest comatose, EP role#

flowchart TD
  Coma[Post-arrest comatose 72 h] --> Exam{Clinical exam}
  Exam -->|Brainstem reflexes absent| BD[Brain-death workup; BAER, apnoea, CTA / EEG]
  Exam -->|Comatose, some reflexes| SSEP[Median-nerve SSEP]
  SSEP -->|N20 absent bilaterally, clean recording| Cat[Catastrophic cortical injury; very poor prognosis]
  SSEP -->|N20 present| Continue[Continue multimodal eval; cEEG, NSE, MRI]
  Cat --> Triangulate[Triangulate with cEEG, NPI, MRI]
  Continue --> Triangulate

8. Try it: interactive widget#

9. Management: how EP findings change decisions#

9.1 Post-cardiac-arrest prognostication#

The 72-hour timepoint (after rewarming and sedation washout) is the canonical EP-driven workflow:

1. Stabilise the patient: complete TTM, allow ≥ 24 h off sedation, treat fever and electrolyte derangements.
2. Clinical exam: pupil reactivity, corneal reflexes, motor response, NPI.
3. Record SSEP with adequate technical quality: stimulus to elicit small thumb twitch, ≥ 500 stimuli per recording, replicated.
4. Read N20 in context: present (continue multimodal evaluation); absent bilaterally with clean recording (high-specificity sign for poor outcome).
5. Triangulate: combine SSEP with cEEG (suppression, burst-suppression, malignant patterns), NPI (< 2), NSE (neuron-specific enolase), MRI (DWI changes), and clinical exam.
6. WLST discussion: bilateral absent N20 plus suppressed cEEG plus NPI < 2 plus malignant MRI is the most reliable bedside dataset for predicting catastrophic outcome; the conversation with family is informed, not made, by this dataset. 1 2 3

9.2 Brain-death determination#

BAER may be used as ancillary electrophysiologic evidence of brainstem silence; the formal determination requires the clinical exam and the apnoea test per the World Brain Death Project framework, with ancillary tests (CTA, TCD, EEG, or BAER) used when the clinical exam cannot be completed. 4 5

9.3 Intraoperative monitoring (brief mention)#

SSEPs and MEPs are standard tools in spine surgery, brainstem surgery, and aortic surgery. The ICU role of these intraoperative EPs is limited to the immediate post-operative recording when intraoperative changes raise concern. 9

10. Clinical contexts#

10.1 Post-cardiac-arrest#

The strongest evidence base. Bilateral absent N20 at 72 h post-arrest, with technically clean recordings, has a false-positive rate < 1% for predicting unfavourable neurological outcome (CPC 3–5). The TTM era has not changed this: the test remains highly specific, with sensitivity around 30–50% (many poor-outcome patients have preserved N20). Pediatric evidence supports the same paradigm; THAPCA and AHA recommend EPs in the post-arrest neuroprognostication bundle. 1 2 10 3

10.2 Severe TBI#

EPs in severe TBI predict outcome but less specifically than in cardiac arrest. Bilateral absent N20 in pediatric severe TBI is associated with poor outcome, but false-positive rates are higher than in post-arrest due to focal injury patterns. EPs in TBI are typically used in conjunction with imaging and clinical exam rather than as a standalone prognostic test. 8 11 9

10.3 HIE / neonatal post-arrest#

BAER is used routinely as part of the neuroprognostic workup for moderate-to-severe HIE. Prolonged or absent wave V is associated with worse outcomes; BAER changes lag aEEG and clinical findings. The combined bundle of aEEG, BAER, MRI, and clinical exam is the standard for term neonatal HIE prognosis. 6 7

10.4 Aneurysmal SAH and DCI#

EPs in SAH are research-grade for DCI detection; the most sensitive electrophysiological tool for DCI is qEEG (alpha-delta ratio). SSEPs may identify large MCA territory ischaemia (loss of N20 over the affected hemisphere) but are too coarse for early DCI detection. 12 13 14

10.5 Pediatric arterial ischaemic stroke#

Unilateral absent N20 in a child with large MCA-territory infarct provides electrophysiologic confirmation of cortical territory loss. Used as adjunct to MRI for prognostic discussion in the post-thrombectomy or natural-history setting. 15 16

10.6 Pediatric ECMO#

Routine EPs are uncommon on ECMO; the modality is reserved for prognostic workup after a suspected neurological event (stroke, haemorrhage) detected on imaging or aEEG. 17 18

10.7 Bacterial meningitis / encephalitis#

BAER is used in suspected brainstem encephalitis and in severe meningitis with brainstem involvement; SSEPs are used selectively. The role is to localise brainstem versus cortical dysfunction in the comatose patient. 19 20

10.8 Brain-death determination#

BAER as ancillary testing has limited specificity for brain death (cochlear preservation, isolated wave I, may persist in some brain-dead patients). The AAN / WBDP framework lists CTA, TCD, and conventional EEG as preferred ancillary tests; BAER is supportive but not preferred. 4 5

10.9 Refractory status epilepticus#

EPs are not a primary tool in SE management. Post-SE, EPs may be used in prognostication when prolonged coma follows refractory SE. 21 22

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

12. Setup and technique#

12.1 Equipment#

• EP machine: Cadwell, Natus, Nicolet, or integrated multimodal device with EP module.
• Stimulator: bipolar surface electrodes for median or tibial nerve stimulation; insert earphones for BAER.
• Recording electrodes: gold-cup or disposable hydrogel; impedance < 5 kΩ.
• Trained electrophysiologist: every ICU EP requires a trained operator. Inter-rater reliability for "absent N20" requires technical expertise; the false-positive rate of < 1% is contingent on this.

12.2 SSEP placement: 6-step protocol#

1. Skin prep at recording sites: C3' or C4' (2 cm posterior to C3/C4), C7 (cervical), Erb's point (supraclavicular fossa), reference (Fz), ground.
2. Apply electrodes: gold cup with conductive paste, secured with collodion or hydrogel pads.
3. Verify impedance: < 5 kΩ at all sites.
4. Stimulator placement: bipolar electrodes over the median nerve at the wrist, cathode proximal. Verify thumb twitch with 5-mA test pulse.
5. Record: 500–2000 stimuli at 4 Hz; replicate with a second independent run; band-pass 30–3000 Hz.
6. Interpret: identify N9, N13, N20 latencies and amplitudes; compare with age-appropriate norms; document reproducibility.

12.3 BAER placement: 6-step protocol#

1. Skin prep at Cz (vertex), ipsilateral mastoid (A1 or A2), ground (Fpz).
2. Apply electrodes: gold cup with conductive paste; impedance < 5 kΩ.
3. Otoscopic exam: clear external auditory canal of cerumen; check tympanic membrane.
4. Insert earphone: deliver clicks monaurally at 80 dB nHL, contralateral ear masked with 65 dB nHL white noise.
5. Record: 2000 stimuli at 11.1 Hz; band-pass 100–3000 Hz; repeat for the other ear.
6. Interpret: identify waves I, III, V latencies and interpeak intervals; compare with age-appropriate norms.

12.4 Quality assurance#

• Replication: every EP recording is replicated with a second independent run; consistent peak presence across replications is required to call a wave "present".
• Stimulus intensity: titrate to elicit small thumb twitch (SSEP) or stable wave I (BAER).
• Temperature: hypothermia prolongs all latencies; record patient temperature at the time of test.
• Sedation: SSEPs and BAERs are robust to sedation; high-dose barbiturate or volatile anaesthetic can flatten responses.

12.5 Common technical pitfalls#

• Lead-off at the cortical site produces "absent N20" that is technical, not clinical.
• Stimulus too weak: produces small or absent N9; re-titrate.
• High impedance: introduces 50/60-Hz mains contamination.
• Patient movement / shivering: contaminates the average; pause and re-record.
• Ear obstruction for BAER: cerumen or middle-ear effusion produces "absent waves" that are conductive, not central.

12.6 Documentation#

The EP report should include: indication, technique (stimulus parameters, recording montage, number of averages, replication), patient temperature and sedation regimen, latency and amplitude of each peak, comparison with age-appropriate norms, interpretation, and a statement of recording quality. Without these the report cannot be acted on confidently.

13. Pitfalls#

• Technical "absent N20": lead-off, high impedance, weak stimulus, or movement can produce false absence. Replication and quality assurance are non-negotiable.
• Sedation effect: SSEPs are robust to most ICU sedation, but very high-dose barbiturate can flatten cortical responses; BAERs are even more robust.
• Hypothermia prolongs latencies: a cooled patient's N20 latency runs longer; correct for temperature in interpretation.
• Pediatric latency tables: using adult cutoffs in young children gives false-positive prolongation; always use age-appropriate norms.
• N20 may persist after catastrophic injury: false-negative rate is high; preserved N20 does not exclude poor outcome.
• Single time-point can be misleading: trends over 24–72 h are more informative than a single recording.
• Cochlear / conductive hearing loss produces absent BAER waves of peripheral origin; not a brainstem finding.
• Brachial plexus injury produces absent N9; the SSEP is uninterpretable for cortical questions.
• Inter-observer variability in EP interpretation is real; the test should be read by a trained electrophysiologist.
• Over-reliance on a single test: the post-arrest prognostic call combines SSEP, cEEG, NPI, NSE, MRI, and clinical exam; no single test should drive WLST.

14. Combine with…#

• Continuous EEG / cEEG: cortical electrophysiology plus pathway-specific EP.
• Amplitude-integrated EEG: bedside cortical trend plus EP for HIE.
• Pupillometry / NPI: brainstem reflex measure plus EP brainstem read.
• TCD: cerebral perfusion plus electrophysiological function.
• Foundations: pediatric physiology: why pediatric latencies differ.
• Integration: post-arrest discordance triage: when EP, NPI, cEEG, and exam disagree.
• Integration: brain death and ancillary testing: role of EPs as supportive ancillary tests.

15. Evidence summary#

16. Recent literature (2022–2025)#

• Amorim 2022 (review): contemporary review of SSEP and EEG in post-arrest prognostication; bilateral absent N20 remains the highest-specificity bedside marker. 1
• Topjian 2021 AHA pediatric: SSEP recommended at 72 h as part of the post-arrest neuroprognostic bundle in pediatric patients. 2
• Naim 2023 (pediatric brain injury post-arrest review): integrates EPs into the broader post-arrest multimodal monitoring stack. 3
• Pediatric MMM consensus (Figaji 2025): positions EPs as tier-2 modality with prognostic role in post-arrest and brain-death contexts. 23
• Brain-death framework (Greer 2020 World Brain Death Project): clarifies the role of ancillary tests including BAER; emphasises clinical exam plus apnoea test as primary. 4
• Tasker 2023 (pediatric neurocritical care review): lists EPs as a tier-2 modality with HIE, post-arrest, and brain-death applications. 27

17. Self-check#