The EEG is not a recording of action potentials.
It is a recording of synapses.

The minimum high-yield framework

1. Origin: summed excitatory and inhibitory postsynaptic potentials in the apical dendrites of cortical pyramidal cells — not action potentials. They are longer, so they summate; the cells are aligned perpendicular to the surface, so their fields add.
2. Amplitude requires synchrony. Desynchronised cortical activity gives a low-amplitude, high-frequency trace; synchronised activity gives a high-amplitude, low-frequency one.
3. Bands — δ 0.5-4, θ 4-7, α 8-13, β 13-30, γ 30-45 Hz.
4. The anaesthetic sequence: β activation → α slowing and anteriorisation → θ → δ → burst suppression → isoelectric.
5. Burst suppression — alternating high-voltage bursts and isoelectric periods. Produced by deep anaesthesia, hypothermia and severe ischaemia.
6. Ischaemic thresholds: EEG slows below a cerebral blood flow of about 20 mL/100 g/min and becomes isoelectric below about 15 mL/100 g/min, while membrane failure occurs only below about 6. That gap is the monitoring window.
7. Evoked potentials: SSEP tests the dorsal columns and is anaesthetic- tolerant; MEP tests the corticospinal tract and is exquisitely sensitive; BAEP is the most resistant of all.

1. Postsynaptic potentials, not action potentials. An action potential lasts about 1 ms; a postsynaptic potential lasts tens of milliseconds. Only the longer event lasts long enough for many cells’ contributions to overlap in time and summate into a measurable field.
2. Geometric alignment. Cortical pyramidal cells are arranged with their apical dendrites all perpendicular to the cortical surface, so each cell’s dipole points the same way and the fields add rather than cancel. Nuclei whose cells are randomly oriented produce no scalp signal, which is why the EEG is a cortical measurement.
3. Synchrony. Because the signal is a sum, amplitude depends on how many cells are doing the same thing at the same time. This is the single most useful idea in the topic: amplitude and frequency move in opposite directions. Synchronised cortex — sleep, deep anaesthesia — gives slow, high-voltage waves. Desynchronised cortex — alert wakefulness, REM sleep, ketamine — gives fast, low-voltage activity.

Two additional patterns are worth naming because they are described rather than measured. Sleep spindles are 11-16 Hz bursts lasting about a second, generated by the thalamic reticular nucleus and characteristic of non-REM stage 2. K complexes are large biphasic transients, also defining stage 2, and can be evoked by an external stimulus — the EEG signature of the sleeping brain registering, and suppressing, a disturbance.

Agent-specific departures from this sequence are examinable in their own right. Ketamine produces high-frequency, low-amplitude activity rather than slowing. Nitrous oxide produces high-frequency activity and characteristically raises processed EEG indices. Sevoflurane and enflurane at high concentration, especially with hypocapnia, can produce epileptiform activity. Opioids at analgesic doses have little effect on the EEG, and dexmedetomidine produces a pattern resembling natural non-REM sleep, with spindles.

The gap between the flow at which the EEG changes and the flow at which cells die is what gives the EEG its value: it provides warning while the tissue is still salvageable. That is the rationale for EEG monitoring during carotid endarterectomy, where a change on cross-clamping indicates inadequate collateral flow and prompts shunting.

Temperature. Hypothermia slows the EEG progressively, with burst suppression at around 24-28 °C and an isoelectric trace at around 18-20 °C. Because hypothermia suppresses both components of cerebral metabolism, an isoelectric EEG under deep hypothermia does not mark the end of protection in the way an anaesthetic-induced one does. Hypoglycaemia, hypercapnia, hepatic encephalopathy and hypothyroidism all slow the trace as well, and none of these can be distinguished from anaesthetic depth by the waveform alone.

An evoked potential is the cortical or subcortical electrical response to a specific, repeated stimulus. Because the response is of the order of 1 µV against a background EEG of tens of microvolts, it is recovered by signal averaging: the response is time-locked to the stimulus and so adds with repetition, while the random background averages towards zero. The signal to noise ratio improves with the square root of the number of repetitions, which is why hundreds or thousands of stimuli are needed and why the technique cannot respond instantaneously.

Two measurements are reported: latency, the time from stimulus to response, and amplitude. The conventional alarm criterion is a 50% fall in amplitude or a 10% increase in latency.

Non-anaesthetic factors that alter evoked potentials, and must be excluded before a change is attributed to surgery: hypothermia (increases latency), hypotension, hypoxaemia, anaemia, hypocapnia, and limb ischaemia from positioning.

Previously examinedOctober 2020 — the electroencephalogram, its origin and the factors that alter it. Worked answers in the library