EEG: Basic Principles and Overview

Introduction

• EEG records the brain’s spontaneous electrical activity, which can be measured through scalp electrodes, offering a continuous and noninvasive measure of network connectivity and brain state.
• It has become an indispensable modality for studying brain function, diagnosing disease, and monitoring anesthesia and sleep. qEEG converts raw waveforms into spectral indices, which help describe and visualize frequency, amplitude, and neuronal complexity, and facilitates the interpretation of EEG.¹

Neurophysiological Basis of EEG

• EEG potentials originate mainly from excitatory and inhibitory postsynaptic currents generated by cortical pyramidal neurons. When many neurons are activated within a narrow timeframe, their current dipoles summate, producing measurable voltage fluctuations.
• The rhythmicity, or oscillatory patterns of the EEG, reflects the synchronized activity of neuronal circuits that connect different brain regions. For example, thalamic pacemaker neurons synchronize cortical firing to generate alpha and sleep-spindle rhythms, whereas cortico-cortical connectivity manifests as higher-frequency beta–gamma oscillations.
• The firing rate and synchronization patterns within such networks determine EEG parameters such as frequency, amplitude, and coherence.^2,3

Raw EEG Patterns, Frequency Bands, Burst Suppression

• EEG activity is conventionally divided into frequency bands (Figure 1). The raw EEG signal represents the summated electrical activity of a range of neuronal populations and networks that oscillate with different frequencies.

• The predominant frequency band measured over certain brain regions correlates with characteristic behavioral states such as wakefulness, meditation, or the different stages of natural sleep.⁴
  • Delta (0.1–4 Hz): Prominent in deep non-REM sleep and coma, reflecting synchronized slow transitions between on-off states of cortical neurons.
  • Theta (4–8 Hz): Seen in drowsiness, early sleep, and meditative states.
  • Alpha (8–12 Hz): Characteristic of relaxed wakefulness with eyes closed, generated by occipital thalamocortical oscillations, found over the occipital region.
  • Beta (12–30 Hz): Dominant during wakefulness and sensorimotor activity.
  • Gamma (>30 Hz): Associated with cognitive engagement and attentional processing.
• Dynamic shifts across these bands represent transitions between alertness, drowsiness, sleep, and anesthesia, or metabolic disturbance.

Burst Suppression and Isoelectric EEG

• Profound depression of neuronal activity may progress to episodes of complete cessation of neuronal firing across some or all brain regions, which manifest as isoelectric epochs on the EEG, interspersed with periods of neuronal activity known as bursts. This alternating pattern of electrical silence and bursts of activity is called burst-suppression (Figure 2).
• With progressive neuronal depression, electrical silence episodes get longer. The ratio of suppressed epochs to bursts is known as the suppression ratio (SR). When the SR approaches 100% long episodes of isoelectric EEG may be present. Burst suppression pattern can be seen in profound hypothermia or overdose with GABAergic anesthetic agents.^3,5–7

EEG Signal Processing and qEEG

• The summated raw electrical signal is typically decomposed into frequency bands and spectrograms using digital signal processing techniques to derive qEEG parameters.^3,6 Commonly used qEEG parameters include:
  • Power spectral density (power spectrum): derived using the Fast Fourier Transform, which is a mathematical method used to convert time-domain signals (brain wave voltage changes over time) into the frequency domain, where the power or average amplitude of individual frequencies is displayed^8,9 (Figure 2).

• Spectrogram displays changes in the power spectrum over time using color-coded heat maps. Higher powers are assigned warmer colors (deeper red), and lower powers are assigned colder colors (deeper blue)^7,8,10 (see OA summary: EEG monitoring and Anesthesia. Link).
• SEF, typically most commonly SEF95, represents the frequency below which ninety-five percent of the total EEG power is contained. When SEF decreases, it reflects cortical slowing, which can occur with sedation, ischemia, or other causes of reduced neuronal activity. When SEF increases, it typically indicates an increase in cortical activation or arousal.¹⁰
• Changes in spectral composition that correspond to physiologic or pathologic brain states are more readily detected, allowing for longitudinal comparison.

EEG and Metabolic Derangement

• EEG is highly sensitive to systemic disturbances of oxygenation, temperature, and metabolism.^1,7,10
  • Hypoxia: Causes progressive slowing from alpha–beta to delta, then voltage attenuation and eventual isoelectric silence
  • Hypercarbia: Produces diffuse slowing via acidosis and cerebral vasodilation; hypocarbia may yield fast, low-amplitude rhythms due to vasoconstriction.
  • Hypoglycemia: Leads to diffuse slowing and sometimes transient epileptiform activity.
  • Hypothermia: Progressively decreases frequency and amplitude; alpha disappears around 30°C, and near-electrical silence occurs near 20°C.
• Because these signatures precede irreversible injury, the EEG serves as an early warning tool in critical care and intraoperative monitoring.

EEG Patterns of Natural Sleep

• Normal sleep is characterized by cyclic transitions between non-REM (non-rapid eye movement) (N1–N3) and REM stages, which have distinct EEG characteristics.⁴
  • Stage N1: Low-voltage mixed-frequency activity as higher frequencies wane.
  • Stage N2: Appearance of sleep spindles (12–15 Hz) and K-complexes, generated by prominent thalamocortical oscillations.
  • Stage N3: Dominated by high-amplitude delta waves indicating widespread neuronal slowing and synchrony.
  • REM: EEG during REM sleep resembles wakefulness, characterized by low-amplitude, mixed-frequency activity and rapid eye movements. It occurs with near-complete muscle atonia and is the stage most strongly associated with vivid dreaming.
• A hypnogram plots these stages over time, showing ~90-minute cycles repeating four to six times nightly. Several anesthetic states mimic non-rapid eye movement features, particularly the delta and spindle patterns produced by GABAergic drugs⁴ (Figure 3).

EEG and Seizure Activity

• Seizures reflect abrupt and excessive synchronization of cortical neuronal firing. They may arise from a focal cortical region or involve both hemispheres from the outset. Focal seizures typically exhibit rhythmic spikes or sharp waves confined to a single area of the recording, and their frequency and morphology may evolve as the seizure progresses. Generalized seizures produce bilateral synchronous spike-and-wave patterns, typically occurring at a frequency of approximately three cycles per second.⁵
• During status epilepticus, high-amplitude rhythmic activity persists with little or no return to baseline between discharges, and the waveform may progress from clearly spiking activity to more subtle rhythmic or periodic patterns as neuronal energy reserves become exhausted.
• In the operating room, seizure recognition can be challenging because frontal montages may not capture ictal discharges arising from posterior or lateral regions. Anesthetic drugs also modify EEG expression, sometimes suppressing overt spiking while leaving rhythmic ictal activity detectable on quantitative displays.⁵ These methods are increasingly used in critical care and intraoperative settings to improve recognition of early or subtle ictal patterns.^1,6