EEG Explained: How Brain Activity & Action Potentials Create Electroencephalograms | Neurological Diagnostics

Focal epileptiform discharges indicate seizure capability and serve as diagnostic/prognostic markers in epilepsy, appearing as synchronous abnormal activity in cortical regions. Generalized epileptiform discharges involve widespread brain involvement, exemplified by 3 Hz generalized spike-wave patterns in childhood absence epilepsy. Diagnostic yield varies: only ~50% of first seizures show abnormalities within 24 hours, increasing to 80-90% with prolonged monitoring. Detection depends on two factors: sufficient cortical involvement (10-20 cm²) and superficial location—deeper structures like mesial temporal regions may escape detection. Understanding these principles guides clinical decision-making in neurological diagnosis.

This video presents five critical epileptiform patterns that clinicians must recognize: (1) Three-per-second generalized spike-wave discharges in absence epilepsy, which present as brief staring episodes in children and require specialist referral since carbamazepine can worsen the condition; (2) Lateralized periodic discharges (formerly called 'plaques'), which indicate serious underlying conditions like subacute strokes, HSV encephalitis, or brain abscess and have a 50-70% chance of being associated with electrographic seizures; (3) Hypsarrhythmia, a severely disorganized EEG pattern with high-amplitude multifocal sharp waves seen in infantile spasms between ages 6-12 months; (4) Centrotemporal spikes, associated with self-limited focal epilepsy of childhood (benign rolandic epilepsy), causing facial paresthesias that children typically outgrow; and (5) Photoparoxysmal response, where intermittent photic stimulation triggers spike-and-wave discharges, strongly associated with generalized epilepsies like juvenile myoclonic epilepsy. Recognizing these patterns is essential for proper diagnosis and treatment planning.

Spike and wave discharges represent abnormal epileptiform activity. Spikes are sharp, transient electrical events; when spikes have multiple phases, they are called poly spikes. These discharges are not normal and indicate pathological brain activity. Generalized spike and wave discharges appearing in both hemispheres suggest generalized epilepsy. To identify lateralization, technologists may expand the time base to 60 or 120 mm/s to see if spikes occur first in one hemisphere before the other. The distinction between symptomatic generalized epilepsy (slow background with spikes) and primary generalized epilepsy (normal background with spikes) is clinically significant for diagnosis and treatment planning.

EEG activation procedures reveal latent abnormalities through eye opening/closing, photic stimulation, hyperventilation, and sleep deprivation. These can reveal epileptiform discharges, asymmetries, and focal abnormalities. Epileptiform discharges include spike-wave complexes, sharp waves, and polyspikes, classified as focal (localized pathology) or generalized (diffuse involvement). About 2% of the population show epileptiform abnormalities without clinical seizures. EEG classification includes idiopathic generalized epilepsy (normal background with discharges), symptomatic focal epilepsy (abnormal background with focal discharges), and symptomatic generalized epilepsy (abnormal background with generalized discharges).

Epileptiform activities are defined by clear beginning and end, detaching from background by frequency, amplitude, and morphology (at least twice background mean). They can be focal, hemispheric, or generalized. Spikes are sharp waves (20-70 ms) with descending phases; sharp waves are slower. Spike-wave complexes are characteristic of generalized epilepsy. Epileptic absence patterns show 3 Hz discharges during hyperventilation. Myoclonic spike patterns show sharp waves followed by slow waves. Pseudo-periodic patterns indicate cortical dysplasia. Phase reversal on bipolar montages localizes seizure generators.

EEG preprocessing involves cascading notch filters to remove power line interference (50 Hz in Europe) and band-pass filters (0.5-30 Hz) to limit bandwidth. A cascading notch filter with second-order configuration and maximum cascading frequency of 100 Hz effectively removes power line hum. Band-pass filters remove DC offsets and limit the frequency range for surface EEG signals. Artifact identification includes EOG artifacts from eye blinks visible on channels near the eyes, and EMG artifacts from muscle activity (clenching teeth) producing high-frequency, high-amplitude signals. Genuine alpha activity appears as low-frequency, high-amplitude oscillations when closing eyes, representing relaxation states.

Common artifacts include eye blinks, eye movements, saccades, muscle tension (buzzy artifacts), respiration (slow rolling artifacts), and cardiac activity. Artifact removal methods include signal space projection (which projects data onto orthogonal spaces excluding artifact components) and Independent Component Analysis (ICA), which assumes statistical independence of sources. ICA can effectively remove cardiac artifacts but requires careful interpretation since it separates mixed signals into independent components that may represent actual neural activity or artifacts.

EEG recordings contain various artifacts that contaminate neural signals. Biological artifacts include eye blinks (most common, detectable by electrodes near eyes), eye movements (create simultaneous positive/negative deflections in lateral electrodes), and muscle contractions (jaw movements, chewing, facial tension produce 60 Hz interference). Cardiac artifacts appear as regular patterns synchronized with heart rate. External interference from power lines (50-60 Hz) creates rhythmic contamination. Electrode movement artifacts produce excessive low-frequency noise and signal distortion. Four main filter types enhance EEG signals: high-pass filters remove slow drifts by allowing frequencies above a cutoff to pass; low-pass filters eliminate higher frequencies since most neural signals fall below 200 Hz; band-pass filters target specific frequency ranges like alpha waves (8-12 Hz); band-reject (notch) filters block specific frequencies (50-60 Hz) to remove power line interference. These techniques enable researchers to isolate signals of interest while eliminating contaminants, improving signal quality for meaningful neural analysis.

Independent Component Analysis (ICA) is a blind source separation technique that decomposes mixed EEG signals into their underlying independent components, allowing researchers to identify and remove specific artifacts like eye blinks by isolating the corresponding component and subtracting it from the data, thereby preserving the original brain signal while eliminating unwanted noise.

High-pass filtering removes slow drift from EEG data caused by environmental factors like gel conductivity changes and skin sweating. A cutoff of 0.1-2 Hz effectively removes these artifacts while preserving brain signals. Common EEG artifacts include eye movements (blinks and saccades), muscle movements from face/jaw/neck, cardiac artifacts, line noise (50/60 Hz from appliances), and cable artifacts. Eye artifacts are among the most prominent because they produce signals much larger than brain signals. The ICA method uses reference electrodes over the eyes to capture eye movement signals and adaptively removes these artifacts by identifying and subtracting eye-related components.

EEG has low spatial resolution (1-2 cm) for localizing brain activity sources but high temporal resolution (millisecond scale, up to 1000 Hz sampling). Techniques with higher spatial resolution (like MEG at millimeter scale) sacrifice temporal precision. fMRI offers high spatial resolution (~1 mm) but low temporal resolution (one map per second). This trade-off between spatial and temporal resolution is fundamental to understanding EEG's strengths and limitations in studying brain function.

EEG offers millisecond temporal resolution (sampling 250-2000 Hz clinically, up to 20 kHz in research), unmatched by fMRI/PET/CT (seconds-minutes). However, EEG has poor spatial resolution compared to fMRI/PET/MRI. EEG detects electrical activity directly while fMRI/PET record blood flow/metabolic changes as indirect markers. Simultaneous EEG-fMRI recordings overcome technical challenges including ballistocardiographic and MRI gradient artifacts. Combined EEG-MEG leverages complementary strengths: MEG provides better surface accuracy while EEG penetrates deeper, allowing correction of inverse problems in source localization. Neither EEG nor MEG detects activity below several centimeters effectively alone, but combined they enable better analysis of deeper brain signals.

EEG offers significant advantages over other neuroimaging techniques despite spatial resolution limitations. Hardware costs are substantially lower than fMRI, PET, SPECT, or MEG systems requiring bulky equipment. EEG achieves millisecond temporal resolution (250-2000 Hz sampling, up to 20 kHz), matching only MEG among non-invasive techniques. It is silent, tolerates subject movement better than most modalities, and avoids claustrophobia and radiation exposure. However, EEG primarily detects postsynaptic potentials in superficial cortical layers, with signals from deep structures, sulci, and midline regions contributing minimally. Axonal action potentials are difficult to detect due to their rapid nature. Multi-modal integration addresses these limitations: simultaneous EEG-fMRI combines temporal and spatial resolution but requires overcoming magnetic field artifacts. Combined EEG-MEG leverages complementary signal origins (extracellular dendritic currents for EEG, intracellular currents for MEG), with simultaneous analysis correcting estimation errors. EEG-PET combinations reveal associations between electrical activity and metabolic processes, valuable for studying pharmacological effects on brain function.

Three primary neuroimaging techniques measure brain activity differently: EEG/MEG detect electromagnetic fields from aligned neuronal firing, requiring spatial synchronization of cortical neurons; fMRI measures blood flow changes as proxies for neuronal activity, providing slower but spatially detailed maps; intracranial recordings offer direct access to brain signals but require surgical implantation. Each technology has distinct temporal and spatial resolution trade-offs, with EEG/MEG capturing rapid dynamics while fMRI reveals broader geographic patterns of activation.

Electroencephalography (EEG) measures brain activity by placing electrodes on the scalp to record time series data, providing good temporal resolution but limited spatial precision. Functional Magnetic Resonance Imaging (fMRI) offers greater spatial resolution by mapping brain activity with detailed location information, but sacrifices temporal resolution as each image requires several seconds to process. fMRI devices are also larger, less practical, and more expensive than EEG equipment.