Authors: Rajan et al.
Journal of Neurosurgical Anesthesiology, April 2026
Key Points
Processed EEG indices can be useful, but they should not be interpreted in isolation.
Discordance between the numerical EEG index and the patient’s true anesthetic state may occur because of artifacts, patient-specific neurophysiology, brain pathology, aging, burst suppression, or drug effects.
Raw EEG and density spectral array interpretation can help clinicians identify misleading index values.
Common artifacts include line noise, electrocautery, EKG contamination, EMG activity, movement, electrode problems, and neurophysiologic monitoring stimulation.
A more complete EEG interpretation may help avoid anesthetic overdosing, underdosing, delayed emergence, awareness, delirium, and cognitive dysfunction.
Summary
This review article explains why anesthesiologists should look beyond processed EEG index numbers and incorporate raw EEG and density spectral array interpretation into intraoperative monitoring. Processed EEG monitors are commonly used to estimate anesthetic depth, but the authors emphasize that these values are generated by proprietary algorithms and can be misleading when artifacts, patient factors, or drug effects distort the signal.
The article begins by noting that intraoperative EEG use has expanded beyond awareness prevention. It is increasingly used to help avoid burst suppression, guide anesthetic dosing in elderly patients, reduce hypotension related to excessive anesthetic depth, and possibly mitigate postoperative delirium and cognitive dysfunction. It may also help identify cerebral ischemia in selected surgical settings.
Commercial monitors such as BIS, entropy, patient state index, narcotrend, and conox provide numerical estimates of anesthetic depth. These numbers are convenient, but they can create false confidence. A high number may be interpreted as inadequate anesthesia, while a low number may be interpreted as deep anesthesia. However, those interpretations may be wrong if the index is affected by artifact, unusual neurophysiology, or drug-specific EEG patterns.
The authors emphasize that many algorithms were developed using relatively narrow patient populations, often healthy younger adults. This limits their reliability in everyday anesthesia practice, where patients may be elderly, neurologically impaired, cognitively vulnerable, medically complex, or undergoing neurosurgery. In these patients, relying only on processed numbers may lead to inappropriate anesthetic dosing.
The article divides causes of EEG index discordance into non-neural artifacts and physiologic or pharmacologic factors. Non-neural artifacts include line noise, electrocautery, EKG interference, EMG activity, patient movement, surgical manipulation, poor electrode contact, and stimulation from neurophysiologic monitoring. These artifacts can falsely increase or distort the EEG index and may make the raw EEG difficult to interpret.
Line noise is a common operating room problem caused by nearby electrical equipment. It may appear as fixed-frequency horizontal lines on the density spectral array and can falsely increase processed index values. If clinicians respond only to the high index by increasing anesthetic dose, they may inadvertently overdose the patient. The authors recommend checking the clinical context, anesthetic delivery, DSA pattern, electrical interference sources, and electrode impedance before changing anesthetic depth.
Electrocautery artifact can create sharp, high-amplitude spikes on raw EEG and broad power increases on the spectrogram. During prolonged electrocautery, there may be too little clean EEG signal for the processed index to be reliable. In those situations, anesthesiologists should avoid overreacting to the numerical index until clean raw EEG and DSA segments are available.
EKG contamination can appear as rhythmic EEG waveforms time-locked to the QRS complex. EMG activity can also contaminate the EEG, especially in higher frequency ranges. EMG may represent artifact, but it may also be clinically meaningful because facial muscle activity or grimacing can indicate inadequate analgesia or inadequate anesthetic depth. The authors emphasize that recognizing the source of the signal matters before deciding how to respond.
The article also discusses movement artifacts, eye blinks, electrode problems, and stimulation artifacts from intraoperative neurophysiologic monitoring. These can create transient high-amplitude patterns or spike-like changes on the EEG. Communication with the neurophysiology team and attention to timing can help clinicians correctly identify these artifacts rather than misinterpreting them as changes in anesthetic depth or seizure activity.
Physiologic factors are equally important. Aging changes the EEG by reducing amplitude and altering frequency patterns. Older patients may have lower EEG power and may be more likely to develop burst suppression at lower anesthetic doses. Processed indices may therefore suggest a lighter anesthetic state than is truly present. The authors suggest adjusting the EEG display scale in elderly patients so low-amplitude raw EEG activity can be better visualized.
Seizures and asymmetric EEG patterns are also discussed. Intraoperative seizures are rare but can occur, and processed frontal EEG monitors may miss activity outside the monitored regions. Asymmetric EEG may occur because of brain lesions, cerebral hemorrhage, infarction, positioning, surgical stimulation, or technical issues such as poor electrode contact. A single index number may not adequately represent these asymmetric or focal processes.
Burst suppression is another important source of discordance. During burst suppression, processed indices may overestimate or underestimate anesthetic depth because the algorithm averages alternating bursts and suppression periods. If burst suppression is unintended, the anesthesiologist should evaluate hemodynamics, drug dosing, clinical context, and anesthetic depth, and consider reducing anesthetic administration when appropriate.
The review also describes pharmacologic paradoxes. GABAergic anesthetics such as propofol and volatile agents usually produce predictable slowing, delta activity, and alpha oscillations. However, ketamine can increase beta and gamma activity and paradoxically raise processed EEG index values even when the patient is adequately anesthetized. Methadone, nitrous oxide, xenon, and dexmedetomidine can also produce EEG patterns that may not fit standard processed-index assumptions.
The authors conclude that clinical decision-making should be based on an integrated interpretation of raw EEG, density spectral array, processed index values, anesthetic drug delivery, cardiovascular status, surgical stimulation, and patient-specific factors. They also suggest that artificial intelligence and individualized baseline EEG calibration may eventually improve interpretation, especially in patients with age-related or disease-related EEG abnormalities.
What You Should Know
This article is a practical reminder that the EEG index number is not the same thing as the patient’s brain state. A processed number can be helpful, but it can also be wrong.
For anesthesia providers, the key takeaway is that raw EEG and density spectral array should be part of routine interpretation whenever EEG monitoring is used. This is especially important in elderly patients, neurosurgical patients, patients with brain pathology, and cases involving ketamine, methadone, dexmedetomidine, electrocautery, EMG activity, or neurophysiologic monitoring.
The clinical risk is real. If a falsely high index is treated as inadequate anesthesia, the patient may receive excessive anesthetic and develop hypotension, delayed emergence, burst suppression, delirium, or cognitive dysfunction. If a falsely low index is trusted too much, the patient may be underdosed and could be at risk for awareness.
Overall, the article supports more education in intraoperative EEG interpretation. The safest approach is not to abandon processed EEG indices, but to use them as one part of a broader assessment that includes the raw EEG, DSA, artifacts, patient context, and pharmacology.
Thank you to the Journal of Neurosurgical Anesthesiology for allowing us to summarize and share this article.