How to read the anesthesia EEG: a six-part series
The most common objection to EEG-guided anesthesia is not that it fails. It is that reading the signal was never part of anyone's training. This series is the remedy.
The six videos below were developed by our co-founders Patrick Purdon and Emery Brown, and each runs about five minutes. The first two build the foundation: why the depth-of-anesthesia indices introduced in the 1990s never became routine practice, and how the EEG waveform and its spectrogram actually work. The remaining four take the major anesthetic agents one at a time, propofol, the inhaled ethers, ketamine, and dexmedetomidine, showing the signature each produces and tracing it back to the receptors and circuits responsible.
The sharpest argument in the series is a pair of failure cases. Ketamine drives fast EEG activity, so a processed index reads a well-anesthetized patient as light.
Dexmedetomidine produces prominent slow-delta activity, so the same index reads a lightly sedated and easily roused patient as adequately unconscious. Opposite errors from a single number, and neither is visible unless you look at the signal underneath it.
CME is available through the associated course.
Part 1: Introduction to the EEG for Anesthesiology
No one reduces the ECG to a single number. Clinicians read the morphology, relate it to what the heart is doing, and act on it. Part 1 asks why the EEG has been treated differently.
The answer it gives is historical. The depth-of-anesthesia indices arrived in the 1990s, before anyone had established the link between the EEG signal and what anesthetic drugs actually do to the brain. Every other physiological monitor works the other way round: understand the physiology first, then build the display.
Five minutes, below, on how that inversion happened, what it cost, and what the spectrogram recovers.
Part 2: The EEG Waveform and Spectrogram
An index and a spectrogram are not two versions of the same thing. The index discards the signal and hands back a number. The spectrogram keeps all of it and simply plots it differently.
Part 2 shows how. It starts underneath the electrode, with the postsynaptic potentials of aligned cortical neurons summing into something large enough to read at the scalp, then builds up to what an oscillation is, how frequency is counted, and how stacking one spectrum after another produces the display already sitting on your monitor.
It also teaches you to do it by eye. Take one second of waveform, count the cycles, and you have the frequency. The math only does it faster and more precisely.
Part 3: Propofol
Propofol is the pattern to learn first, because everything after it is described as a comparison. Slow-delta plus alpha, and once you have seen it you will see it in most of your cases.
Part 3 explains where each half comes from. The alpha is thalamus and cortex still communicating, but squeezed into a narrow band around 8 to 12 Hertz instead of the wide range they use when a patient is awake. The slow-delta marks something different: distant cortical regions falling out of step with each other, so local communication survives while global communication does not.
It also covers induction and emergence, including the zipper-like opening of the alpha band as a patient comes back.
Part 4: Inhaled Ether-derived Anesthetics
If you run inhaled agents, the useful part here is not that the pattern resembles propofol, though at surgical levels it does. It is what happens above MAC and at the end of the case.
Around MAC and above, the alpha drops in frequency and a strong theta oscillation appears, marking a deeper state. Theta is also the first thing to disappear as concentration comes down. That gives you a visible marker for a boundary you currently take from the agent monitor.
The last two minutes follow a real case through the switch from sevoflurane to nitrous oxide at closure, a maneuver most of us perform routinely without ever seeing what it does to the brain.
Part 5: Ketamine
Ketamine produces anesthesia by taking the brakes off rather than by pressing down. It blocks NMDA receptors on inhibitory interneurons, which stops them restraining the excitatory neurons downstream, so activity rises rather than falls.
The EEG follows. Fast oscillations around 25 to 32 Hertz, an active-looking trace on a patient who is anesthetized.
That is the trap. Processed indices treat low frequency as deep, which holds for propofol and sevoflurane and does not hold here. Give ketamine and the number can climb while your patient is exactly where you want them.
Part 6: Dexmedetomidine
Dexmedetomidine is the only agent in this series that produces something close to sleep, and it does so by borrowing the pathway sleep uses. It quiets the locus coeruleus, norepinephrine falls, and the preoptic area is freed to suppress the arousal centers, which is thought to be part of how non-REM sleep begins.
The EEG shows it. At low doses, slow-delta with genuine spindles, the same feature that marks stage 2 sleep. Push the dose and the spindles shrink while the slow waves grow, until the trace resembles slow-wave sleep.
Which sets up the more dangerous of the two index failures. Ketamine makes the number read high on an anesthetized patient. Dexmedetomidine makes it read low on a patient who is sedated, easily roused, and nowhere near a surgical plane.
Comments