Role of Neurophysiologic Testing in Prognostication of Hypoxic-Ischemic Encephalopathy Authors: Erik Westhall, Marjolein Admiraal Published: 2026 ABSTRACT Summary:Electroencephalography (EEG) and somatosensory evoked potentials (SSEP) are important components of multimodal prognostication in comatose patients after cardiac arrest (CA). The EEG changes considerably during the first days after CA. The prognostic value of some EEG patterns is time-dependent, whereas that of others is not. Persistence of burst-suppression with heterogeneous (variable) bursts and generalized suppression (<10 µV) ≥24 hours after CA strongly predicts poor outcome. Burst-suppression with identical or highly epileptiform bursts and periodic discharges on a suppressed background strongly predict poor outcome, also within the first 24 hours during ongoing sedation and temperature management. These synchronous patterns are best assessed with early continuous EEG-monitoring (cEEG), because their prevalence gradually decreases. Early (<12 hours) return of a continuous normal-voltage background activity on cEEG is a strong predictor of good outcome. A reactive and continuous background on a late routine-EEG also indicates a good prognosis. Full-montage or simplified cEEG provides early prognostic information and improves sensitivity for both poor and good outcome prediction compared with a single late routine-EEG. Bilateral absence of cortical potentials (N20) on median nerve SSEP is one of the most reliable tools for predicting poor outcome, including during ongoing sedation. Sensitivity is limited, but specificity is near 100% if recordings are of sufficient quality, with clearly discernible peripheral potentials and low noise levels. Currently not included in guidelines, very low cortical SSEP amplitudes (approximately 0.5 µV or less) are strongly predictive of poor outcome. However, further research is needed to establish consensus criteria. Comatose survivors of out-of-hospital cardiac arrest (CA) represent a common patient group in the intensive care unit (ICU). Approximately half of these patients will survive, and most who survive will have a good long-term outcome. Most deaths occur after withdrawal of intensive care, typically after an anticipated poor neurologic prognosis due to severe hypoxic-ischemic brain injury.1,2 Importantly, neuroprognostication and the decision whether to withdraw treatment should be considered as separate processes, because legislation and practice differ between countries. This topical review discusses the role of neurophysiologic testing in neuroprognostication. The type and circumstances of the CA provide an indication of the prognosis but cannot reliably predict outcome. Present guidelines recommend prolonged observation and the use of multiple tools to assess prognosis.3–5 Clinical examination is the foundation of the prognostic assessment, combined with neurophysiologic methods, neuroimaging, and blood biomarkers of brain injury.3–5 Most of these prognostic signs can be assessed during the first few days in the ICU, including results of neurophysiologic testing. To ensure reliable prediction of poor neurologic outcome, current guidelines state that prognostic information should not be used until after 72 hours, provided that the patient is still in deep coma without confounding factors such as metabolic derangement or sedation.4 Despite the guideline recommendation to delay decisions regarding WLST, neurophysiologic tests can provide reliable information on both good and poor prognosis even during the first few days in the ICU. This early prognostic information may be useful for planning care and for informing relatives. The prognostic value of certain signs, such as pupillary and corneal reflexes, is time-dependent and varies according to the time of assessment.3–5 Some neurophysiologic findings are also time-dependent while others are not. In this review, we describe the prognostic value and clinical application of the two neurophysiologic tests that are most commonly used for prognostication after CA, electroencephalography (EEG), and somatosensory evoked potentials (SSEP).3–6 We acknowledge that most evidence is derived from adult out-of-hospital CA cohorts with presumed cardiac cause. EEG FOR PROGNOSTICATION AFTER CARDIAC ARREST EEG is widely accessible and recommended in guidelines for both prognostication and detection of seizure activity. It is one of few prognostic tools capable of indicating both good and poor prognosis. Standardized Interpretation of EEG EEG has been used for more than 50 years to evaluate prognosis in coma after CA.7 In the past decade, studies have increasingly adopted the standardized EEG terminology published by the American Clinical Neurophysiology Society (ACNS).8 Standardization is crucial for meta-analysis, systematic review, and communication between the electroencephalographer and the referring physician.9–11 The two most important aspects of EEG assessment are included in the ACNS terminology: characterization of the EEG background activity and superimposed rhythmic or periodic discharge patterns. The terminology also provides qualitative definitions of EEG reactivity. In recent years, the EEG terminologies of the International Federation of Clinical Neurophysiology12 and ACNS have been harmonized, resulting in a high degree of international consensus on the definitions of important EEG patterns after CA. Importantly, the recent 2021 version of the ACNS terminology now also includes the burst-suppression subtypes with identical or highly epileptiform bursts.8 Clinical interpretation of EEG is still primarily based on visual qualitative assessment, although quantitative tools, such as various cEEG trends, are gradually being introduced to assist the electroencephalographer. It is well known that interrater variability exists for visual assessment and the ACNS terminology was developed to limit this and improve reproducibility.8 Key EEG predictors of poor outcome include suppression with or without periodic discharges and burst-suppression, including the subtype with identical bursts. These patterns show substantial interrater agreement among experienced EEG readers.13–15 According to the ACNS terminology, the proportion of suppression periods (amplitude <10 µV) in burst-suppression should constitute 50% or more, and in a suppression pattern the entirety of the recording (≥99%) should be suppressed. EEG patterns that are less reliable for prognostication, such as low-voltage background (amplitude 10–20 µV), discontinuous background (10%–49% suppression), or unreactive EEG have lower interrater agreement.13,16 Important prognostic EEG patterns are presented in Fig. 1. Important EEG patterns that predict poor outcome. EEG is shown as simplified EEG in two contralateral channels. Note that synchronous patterns (for instance burst-suppression with identical bursts) and early status epilepticus starting from burst-suppression strongly predict poor outcome even during the first 24 hours and ongoing sedation. Time-dependency of EEG In postanoxic coma, EEG activity changes considerably during the first hours and days after CA.17 During resuscitation and immediately after return of circulation, most patients show a suppressed EEG (<10 µV). In patients with a good outcome, EEG activity returns after a typical sequence from suppression to burst-suppression with heterogeneous (variable) bursts with gradually decreasing duration of suppression periods, to a discontinuous background. Amplitudes also improve gradually and often within the first day a continuous normal-voltage (>20 µV) background activity returns, see Fig. 2. Shows the gradual return of EEG background activity in a patient with good outcome. EEG is shown as simplified EEG in two contralateral channels. Early (<12 hours) return of a continuous normal-voltage background strongly predicts a good neurologic outcome. Note that suppression and burst-suppression with heterogeneous (variable) bursts may occur during the first 24 hours in patients with a good outcome. Considering that the reliability of some EEG patterns is highly time-dependent, we describe the evidence for early (within 24 hours after CA) and late (≥24 hours after CA) EEG for prognostication in separate sections. In each section, we describe the predictive value for both poor and good outcome. Figure 3 summarizes the prognostic value of important time-dependent and time-independent EEG patterns in early (<24 hours) and late (≥24 hours) phases after CA. Prognostic value of important time-dependent and time-independent EEG patterns in the early (<24 hours) and late phase after cardiac arrest (≥24 hours, when sedation is reduced, often corresponding to 2–3 days). Patterns in green are strongly predictive of a good neurologic outcome, patterns in yellow cannot reliably predict a certain outcome, and patterns in red are strongly predictive of a poor outcome, with specificity close to 100% (>95% in nearly all studies). The figure is based on evidence from the literature and reflects the opinion of the authors. Note that an early (<12 hours) return of a continuous normal-voltage background and a late reactive continuous background strongly predict good outcome. Synchronous patterns and early status epilepticus starting from burst-suppression strongly predict poor outcome during the first 24 hours and with ongoing sedation. Early EEG for Prediction of Poor Outcome During the first 24 hours after CA, burst-suppression with identical bursts, burst-suppression with highly epileptiform bursts, and generalized periodic discharges on an otherwise suppressed background (<10 µV)18,19 are reliable predictors of a poor outcome. Although it is well known that EEG activity is affected by sedation in a dose-dependent manner,20 these patterns, also known as “synchronous patterns”, predict poor outcome with 100% specificity at all time points, regardless of ongoing sedation with propofol or midazolam in clinically used doses.18,19,21–25 Their sensitivity ranges from 30% to 50% within 24 hours and gradually decreases to below 10% at 2 to 3 days after CA,19,21 because these patterns are typically transient and disappear within 1 to 2 days after CA.18,19,21,26 Figure 4 shows a patient with early synchronous patterns who had a poor outcome. Shows a patient with early synchronous patterns, which strongly predict poor outcome also during the first 24 hours and ongoing sedation. EEG is shown as simplified EEG in two contralateral channels. The patient had a poor outcome. In very rare cases, burst-suppression with identical bursts might be seen in patients who recover. For example, a case where propofol sedation was combined with ketamine and inhaled sevoflurane,27 and two cases where baclofen was combined with a serotonin-inhibitor at the time of the arrest.28 The prognostic significance of a suppressed background (<10 µV) is time-dependent. Prognosis is uncertain if suppression is seen during the first hours after CA. A suppressed background between 12 and 24 hours is associated with poor outcome, although some patients recover.21,29,30 If generalized suppression persists until 24 hours after CA, no survivors with a good outcome have been described, regardless of ongoing sedation. The exact time point between 12 to 24 hours at which suppression should be considered a reliable predictor of poor outcome can be debated. For clarity in this review, suppression is considered a time-dependent late (≥24 hours) predictor (see below). Approximately one-third of comatose patients after CA show rhythmic and periodic patterns compatible with possible or definite seizure activity, as defined in the ACNS terminology and Salzburg criteria.8,31 From a prognostic point of view, an early start (<24 hours) of possible or definite seizure activity arising from a burst-suppression background predicts poor outcome with 100% specificity.1,2,32,33 Most of these patients also fulfill the criteria of the synchronous patterns.34 However, among patients with a favorable electrographic postanoxic status epilepticus profile, that is, late start (>24 hours) of possible or definite seizure activity from an already established continuous background, approximately one-third achieve good long-term outcome. Patients who recover typically lack signs of extensive brain injury on all other prognostic tests and show improved level of consciousness within the first week.33–40 Rhythmic and periodic patterns are further discussed elsewhere in this issue. Early EEG for Prediction of Good Outcome Good outcome prediction is best performed using early cEEG monitoring to assess the time at which a nearly continuous normal-voltage (>20 µV) background activity returns: the earlier this occurs, the better the neurologic prognosis. A good outcome is most likely if the background is restored within 12 hours, and approximately 90% of these patients achieve a favorable neurologic outcome.19,21,29,32,41–45 Figure 2 presents a patient with early return of a continuous background activity who had a good outcome. Late EEG for Prediction of Poor Outcome Late reliable predictors (≥24 hours after CA) of a poor outcome are heterogeneous (variable) burst-suppression and generalized suppression (<10 µV). These patterns, sometimes grouped as highly malignant patterns, have been extensively studied in the literature,9,10,24,29,46–50 and are included in the prognostic algorithm of the present guidelines.4,9,46 Most studies were performed by experienced EEG readers, and reported specificities are close to 100%. These patterns are typically recorded using a routine-EEG in the time frame of 2 to 3 days after arrest under normothermic conditions. At this late time point, sedation is usually stopped or reduced and these patterns are not statistically affected by ongoing low-dose sedation commonly used in the postanoxic setting.46,51 However, some patients with late heterogeneous burst-suppression but an eventual good outcome (false-positives) have occurred, most likely because of higher dose of sedation used for treatment of postanoxic seizures or invasive procedures.47,52 A recent large multicenter study evaluating the prognostic ability of a late routine-EEG interpreted by local EEG reviewers with varying experience reported a 7% false-positive rate for burst-suppression or suppression with or without periodic discharges.49 Another study showed higher interrater variability among junior EEG reviewers.15 These findings highlight the need for involving EEG readers with varying experience in future studies. The sensitivity of heterogeneous burst-suppression or suppression in the late time period, beyond 24 hours, is typically approximately 30% to 40%,9,50 although some studies report up to 50%,46,49 and remains approximately 30% to 40% during the following days.49,53 This relatively high sensitivity likely reflects selection of patients with prolonged coma. Other patterns such as low-voltage (10–20 µV), discontinuous, unreactive EEG background activity, or periodic discharges on a continuous background are not reliable predictors of a poor outcome, owing to too many false-positives. Combinations of features (e.g., low-voltage and discontinuous EEG) have shown very high specificity for poor outcome prediction late after CA, but require further validation.46,50,54 A routine-EEG for prognostication traditionally includes reactivity testing. Standardized stimulation protocols have been proposed,55 often consisting of sound and pain stimuli, but no guidelines exist for stimulation procedures and interpretation. The absence of EEG reactivity itself is not a reliable predictor of poor outcome.16,49 When suppression or burst-suppression is present, the EEG is typically unreactive or shows only stimulus-induced rhythmic, periodic, or ictal discharges,56,57 or stimulus-induced bursts.58 In the unexpected case of reactive burst-suppression, the number of false-positive patients increase.49 Late EEG for Prediction of Good Outcome Over time, the predictive value of continuous EEG background activity for prediction of a good outcome declines. Patients restoring a continuous background beyond 2 to 3 days after CA have a low chance of recovery.21,41 Thus, without knowledge of the early EEG evolution, the observation of a continuous background late after CA, for instance on a single routine-EEG, has limited prognostic value.19,21,32,42 However, in the late phase, a reactive continuous normal-voltage (>20 µV) background without periodic discharges, is predictive of good outcome, with reported predictive values up to 90%.46,59 Absence of burst-suppression and suppression in the late phase, particularly with preserved EEG reactivity, may indicate good prognosis, although predictive ability is limited.45,49,60 Arguments for cEEG Both routine-EEG and cEEG are used in post-CA care.3,61 Because traditions vary between hospitals and countries and clear evidence is lacking, we discuss both strategies, including the use of simplified cEEG monitoring with a reduced electrode montage. We focus on cEEG strictly from a prognostic perspective and do not consider costs, seizure detection, or therapeutic implications as discussed elsewhere in this issue. A strategy that detects early “synchronous patterns” improves sensitivity of poor outcome prediction compared with a late EEG assessment alone.19,21 Given their transient nature, early cEEG is the most suitable method to detect these patterns. Studies exploring detection of early predictive patterns with routine-EEG are scarce, and their prevalence using this strategy is low.24 Prolonging cEEG allows additional assessment of late EEG predictors, for instance heterogeneous burst-suppression, increasing sensitivity further to approximately 50%.21 Another advantage of cEEG is that when an early suppressed background is detected, monitoring can be continued to check persistence after 24 hours. Continuing cEEG beyond 2 days after CA does not significantly improve sensitivity for poor outcome prediction.19,21 If started within 12 hours, cEEG can also detect early restoration of a continuous normal-voltage background, which strongly predicts good outcome. Arguments for Routine-EEG Routine-EEG performed after the end of temperature management and tapering of sedation has been extensively investigated. Sensitivity for poor outcome prediction of a single late routine-EEG (>24 hours after CA) during office hours is likely lower than strategies including early cEEG, although no direct comparisons exist. A strategy to perform two routine-EEGs, one early and one late, has been investigated in comparison with cEEG in a small randomized trial, and cEEG detected more seizures, but outcomes were comparable.62 Other cohort studies also reported similar prognostic performance when comparing cEEG and routine-EEG.56,63 A late routine-EEG showing a continuous and reactive background predicts good outcome, with only slightly lower sensitivity than early (<12 hours) return of a continuous background on cEEG. EEG Strategy – Opinions of the Authors We acknowledge that there is no clear evidence to support one EEG strategy over another. In medium-sized or large hospitals that regularly treat patients with CA, we suggest performing cEEG monitoring, ideally in all patients, and preferably with an electrode montage that can be started and maintained by staff that is available also outside office hours, for instance by ICU nurses, because EEG technicians on call are commonly lacking. Full-montage cEEG monitoring requires trained technicians, but its availability in postanoxic care varies between centers and health care systems. Studies have shown similar prognostic performance for reduced montages compared with full-montage cEEG,21,32,64–71 see examples in Fig. 5. American guidelines (AHA) also mention the possibility of using a reduced cEEG montage.6 Examples of simplified cEEG montages with a reduced number of electrodes that have been used in the literature, (A) evenly distributed over the scalp (22, 70), or (B) a subhairline montage (75). cEEG monitoring should ideally be started within 12 hours after the arrest. If an early predictor of poor outcome or early status epilepticus arising from burst-suppression is documented, cEEG may be stopped after the first day, especially if other signs of poor prognosis are present, for instance, markedly elevated neuron-specific enolase levels. Similarly, if an early predictor of good outcome appears, especially if the patient shows signs of improved level of consciousness without myoclonus, cEEG may also be stopped early. If cEEG during the first 24 hours is prognostically indeterminate, then cEEG can be continued to determine whether a late predictor such as heterogeneous burst-suppression is detected after 24 hours. If cEEG is indeterminate during the first 24 hours and possible or definite seizure activity subsequently arises from a continuous background later than 24 hours after CA, we suggest prolonging monitoring, because approximately one-third of these patients eventually have a good outcome, provided no signs of extensive brain injury are found using other prognostic tools. Occasionally, background activity may be difficult to assess with a reduced electrode montage, for example, to decide whether the background is suppressed (<10 µV) or low-voltage (10–20 µV). In this scenario, an additional full-montage routine-EEG during office hours may be useful and, if necessary, muscle relaxants can be administered.72 In smaller hospitals that treat patients after CA less frequently, or in resource-limited settings, our opinion is that a late routine-EEG during office hours approximately 2 to 3 days after CA seems the most pragmatic strategy. It can be debated whether a single early (12–24 hours after CA) routine-EEG could be an alternative in limited resource settings to try to detect early predictors of poor or good outcome, but because patterns may be transient and the possibility for performing routine-EEG may be restricted to office hours there is a risk that the findings will be prognostically indeterminate. In settings where cEEG is not used, an additional routine-EEG could be considered if clinical convulsions occur early after CA, as suggested in the guidelines.4 Assessment of the EEG background pattern and the presence of rhythmic or periodic activity may provide prognostic information and potentially guide therapeutic decisions.34,36,37 However, restoration of a (nearly) continuous background before the onset of seizure activity is the strongest marker of a potential good outcome34,37 and can in practice only be documented using cEEG. EEG for Stratification In addition to being a purely prognostic marker, EEG can also be used to optimize clinical decision making by stratifying patients for other prognostic tests and potentially personalizing treatment. For instance, an ongoing randomized trial explores the possibility to abort temperature management and wake the patient if the early cEEG shows a favorable pattern.73 SSEP FOR PROGNOSTICATION AFTER CARDIAC ARREST Median nerve SSEP for prediction of poor outcome is recommended in guidelines.3,4 Bilaterally absent cortical responses in a high-quality recording reliably predict poor outcome, but with limited sensitivity. In recent years, several studies have explored the use of very low-voltage cortical SSEP potentials to predict poor outcome and whether SSEP can be used to predict good outcome. This review will discuss short-latency SSEP potentials, but not later potentials or other promising modalities of evoked potentials. SSEP Recording and Interpretation The median nerve is electrically stimulated at the wrist with a rate of 3 to 5 Hz, resulting in a clear abduction of the thumb. To generate a signal with sufficient signal-to-noise ratio, hundreds of stimuli (commonly around 500) are averaged, and the measurements are repeated at least twice to ensure reproducibility. To confirm absence of N20 waves, it is recommended to record cortical potentials with two different references, for example, Fz and the contralateral parietal cortex.74 A median nerve SSEP recording with preserved cortical potentials is shown in Fig. 6A. A, Example of a median nerve SSEP with bilaterally present N20 cortical potential (one side presented here). Peripheral responses are clearly distinguishable, and noise level is well below 0.25 µV. Different possibilities for calculating N20 amplitude are shown. B, Example of SSEP with very low-voltage N20 cortical potentials (0.5 µV) bilaterally when measured N20-P25 (one side presented here). When the maximum peak-to-peak amplitude is measured, the cortical amplitude is 1.0 µV. This illustrates that a detailed description of the measurement technique is essential when very low-voltage SSEP amplitude is reported. Common ways of measuring amplitude of cortical potentials are baseline-N20, N20-P25 (most used), and highest peak-to-peak amplitude between 4.5 and 50 miliseconds after the stimulus,75–78 see Fig. 6A. The baseline-N20 amplitude is typically considerably lower than the N20-P25 or peak-to-peak amplitudes. To state that cortical responses are absent, a high-quality recording with a low noise level and clearly discernible peripheral responses is crucial. A present N9 potential at Erb's point, followed by the N13 potential recorded from the spinal cord, ensures the integrity of the peripheral sensory system up to the dorsal column. The criteria for present peripheral responses required to consider a SSEP assessment conclusive vary in the literature. Most studies require the presence of both N9 and N13, whereas others only mention N9. Subsequent presence of P14 and N18 recorded on the scalp reflects far field potentials from subcortical structures (brain stem), but only few studies include this as a quality criterion, and they may be absent in patients with severe hypoxic injury. Recording these potentials in a dedicated channel reduces the risk of misinterpreting them as N20 waves.74 SSEP interpretation is subject to interrater variability. Recordings with noise levels exceeding 0.25 μV have been associated with increased interrater variability.79,80 Many studies and laboratories use 0.25 μV as the acceptable upper limit of the noise level. The use of neuromuscular blockade improves the quality of the recording and should always be considered when no clear cortical responses are detected. Absence of SSEP Cortical Potentials In current guidelines, bilaterally absent cortical SSEP potentials (N20) is one of the indicators of poor outcome in comatose patients with CA. Nearly all studies in the literature report 100% specificity for poor outcome, but cases of alleged false-positives exist. False-positive rates might be underestimated in the presence of withdrawal of life-sustaining treatment, resulting in a self-fulfilling prophecy.81 A recent article critically discussed false-positive cases identifying suboptimal recordings or missing clinical information in these case reports.82 The sensitivity for poor outcome prediction of bilaterally absent cortical responses in the general CA population is approximately 30%, which is relatively low compared with other prognostic markers.4 Recent observational studies often report higher sensitivities, but this can be largely attributed to selection bias when primarily examining patients with very prolonged coma. Repeating a SSEP recording when N20 waves were deemed absent in a recording with sufficient quality does not contribute to the prognostic certainty, as absent SSEP responses do not re-emerge in the first five days.83 Amplitude of SSEP Cortical Potentials In recent years, more evidence has emerged for the prognostic value of cortical SSEP wave (N20) amplitudes. Very low-voltage cortical SSEP amplitudes (approximately 0.5 μV or less) are strongly predictive of poor outcome.4,75–77,84,85 However, proposed cutoffs to define very low-voltage cortical amplitudes vary in the literature, ranging from approximately 0.5 to 1.0 μV. Amplitude criteria for cortical SSEP potentials are not yet included in prognostic guidelines,4 because of varying electrode montages, differences in amplitude measurement methods, and insufficient evidence regarding influence of sedatives and timing of the SSEP recording. Combining these factors, bilateral amplitudes of ≤0.5 µV measured as N20-P25 or peak-to-peak (not baseline-N20 amplitude) could be considered a conservative definition of very low-voltage cortical amplitudes, according to the authors of this review. Figure 6B presents a SSEP recording with very low-voltage cortical amplitude. Traditionally, short-latency SSEP has not been used to predict a good outcome, but a few studies have explored the prognostic value of high-amplitude N20 SSEP waves.75,78,86,87 Further research is needed to determine whether cortical SSEP amplitudes above a certain threshold provide additional value for good outcome prediction compared with favorable EEG findings or low levels of blood biomarkers of brain injury. SSEP Timing In guidelines, SSEP is recommended beyond 24 hours from CA. However, early SSEP recordings within 24 hours have been shown to have similarly high specificity.74 The optimal time point and in which patients to perform SSEP can be debated. A common strategy is to perform SSEP in patients with prolonged coma beyond 2 to 3 days after CA if prognosis is indeterminate. When cEEG or a late routine-EEG indicate good prognosis, SSEP will not add prognostic information, as nearly all these patients have preserved cortical SSEP potentials.83,88,89 If multimodal prognostication, including EEG, fulfills criteria for a poor prognosis, a SSEP is often considered redundant. Alternatively, SSEP may be performed as a confirmatory test because of its very high specificity, although sensitivity is limited. SSEP and Sedation Sedation in clinically used doses has not been shown to affect the reliability of absent cortical responses. However, sedation may increase latencies and decrease amplitudes.74,82 Propofol may suppress amplitudes by 10%, whereas midazolam has been reported to reduce amplitudes up to 60%.74 Since sensitivity for prediction of poor outcome with SSEP does not decrease over time, SSEP is ideally performed after the end of temperature management and reduction of sedation. CONCLUSIONS EEG and SSEP are valuable components of multimodal prognostication in comatose patients after CA. With growing knowledge of the temporal development of time-dependent and time-independent EEG patterns and emerging criteria for very low-voltage cortical SSEP potentials, the prognostic utility of these tools has improved. The keys for optimizing the use of neurophysiogic testing in the postanoxic setting are collaboration and clear communication among clinical neurophysiologists, neurologists, and intensivists.