Electroencephalography

Electroencephalography (EEG) is the recording of electrical activity of the brain through electrodes placed on the scalp. It measures the summed postsynaptic potentials of large populations of neurons, filtered through the skull, scalp, and cerebrospinal fluid. The resulting signal is a voltage fluctuation — typically in the microvolt range — that carries information about the brain's state, from the slow oscillations of deep sleep to the fast gamma rhythms of active attention.

The EEG is the oldest non-invasive window into the living brain. Hans Berger recorded the first human EEG in 1924, detecting what he called the "alpha rhythm" — an 8–13 Hz oscillation prominent when a person closes their eyes and relaxes. Berger's discovery was initially dismissed by the scientific establishment, but it established a fundamental principle: the brain's electrical activity is not merely noise; it is structured, rhythmic, and meaningfully correlated with mental states.

The Physics of the EEG Signal

The EEG signal is not a direct measure of neural firing. Individual action potentials are too fast and too spatially localized to be detected at the scalp. Instead, the EEG reflects the synchronous activity of thousands or millions of neurons — primarily pyramidal cells in the cortex, whose dendritic trees are oriented perpendicular to the cortical surface and whose simultaneous postsynaptic currents create detectable dipole fields.

The spatial resolution of EEG is limited by the smoothing effect of the skull, which acts as a low-pass filter for electrical potentials. The temporal resolution, however, is excellent — milliseconds, matching the speed of neural computation. This temporal precision is the EEG's unique advantage over other neuroimaging techniques. Where fMRI measures blood flow with a lag of seconds, and PET measures metabolism with a lag of minutes, the EEG captures neural dynamics in real time.

The EEG signal is typically analyzed in the frequency domain, where its power spectrum reveals characteristic bands: delta (0.5–4 Hz, associated with deep sleep), theta (4–8 Hz, associated with drowsiness and memory encoding), alpha (8–13 Hz, associated with relaxed wakefulness), beta (13–30 Hz, associated with active thinking and motor control), and gamma (30–100 Hz, associated with attention and conscious perception). These bands are not arbitrary categories; they reflect distinct oscillatory mechanisms generated by different neural circuits and neurotransmitter systems.

From Raw Signal to Cognitive Interpretation

The transition from voltage fluctuations to cognitive claims is treacherous. Raw EEG is contaminated by artifacts: eye movements, muscle tension, heartbeat, and electrical interference from the environment. Modern EEG analysis relies on sophisticated preprocessing — independent component analysis, spectral filtering, and spatial filtering — to isolate neural signals from noise.

Beyond preprocessing, the interpretation of EEG requires linking frequency-domain features to cognitive states. The event-related potential (ERP) is a time-domain technique in which EEG is averaged over many trials of a repeated stimulus, revealing a stereotyped sequence of positive and negative deflections (the P300, N400, P600, and others) that correlate with specific cognitive processes. The P300, for instance, is a positive deflection approximately 300 milliseconds after an unexpected stimulus; it is one of the most reliable biomarkers of conscious detection.

In the frequency domain, neural oscillations are analyzed for their relationship to cognitive function. Gamma-band synchronization, first observed by Wolf Singer and Charles Gray in the visual cortex of cats, has been proposed as a mechanism for feature binding — the neural correlate of perceiving a unified object from its separate attributes. Theta oscillations in the hippocampus are coupled to gamma oscillations in a hierarchical manner, suggesting that the brain uses cross-frequency coupling to organize information across multiple timescales.

Clinical and Technological Applications

EEG is indispensable in clinical neurology. It is the primary tool for diagnosing epilepsy, detecting abnormal spike-and-wave discharges that characterize seizure activity. In sleep medicine, EEG is the gold standard for staging sleep and diagnosing disorders. In anesthesia, EEG-derived indices such as the bispectral index monitor depth of unconsciousness during surgery.

The technological frontier of EEG is brain-computer interfaces (BCIs). Because EEG is non-invasive, portable, and relatively inexpensive, it is the most accessible modality for translating neural activity into control signals for external devices. Motor imagery BCIs — in which a user imagines moving their hand, and the resulting sensorimotor rhythm desynchronization is detected and translated into a cursor movement — have enabled paralyzed individuals to communicate and control their environment. The limitation is information bandwidth: EEG BCIs are slow and error-prone compared to invasive alternatives, but they are also safe and scalable.

The EEG as a System-Level Measurement

The EEG is a paradigmatic example of an emergent systems measurement. No single neuron produces the EEG; the signal emerges from the collective dynamics of neural populations. The individual neuron is a binary switch; the EEG is a continuous oscillation. The transition from spike to wave is a transition from microscopic mechanism to macroscopic phenomenon — and it is precisely this transition that makes the EEG cognitively interpretable.

The relationship between EEG and the underlying neural dynamics is still incompletely understood. The Local Field Potential (LFP), recorded directly from within the brain, is the mesoscale link between single-neuron spikes and scalp EEG. LFPs reflect synaptic currents, population firing rates, and glial activity in a local volume of tissue, and they are the primary source of the signals that eventually reach the scalp. Understanding how LFPs scale to EEG — how local synchrony becomes global coherence — is one of the central problems of systems neuroscience.

The EEG's most profound limitation is also its most profound insight: it measures correlation, not causation. An EEG rhythm tells us that a population of neurons is firing in synchrony, but it does not tell us why. The search for causal mechanisms — for the ionic currents, synaptic connections, and network architectures that generate specific rhythms — requires combining EEG with other techniques: computational modeling, invasive recordings, and perturbation studies. EEG alone is a map; to understand the territory, we need the map and the mechanism.

_The EEG is not a photograph of the mind. It is a spectrogram — a reading of the brain's temporal signature. The fact that we can read this signature at all, from outside the skull, is remarkable. The fact that we have not yet learned to read it fluently is a measure of how much we still do not understand about the dynamics of neural populations. The brain does not merely produce oscillations; it thinks in them. The EEG is our imperfect transcript of that thought._