EEG Measures Electrical Activity

Does an EEG detect electrical activity? If so, what is the source of this electrical activity?

Yes. EEG measures tiny voltage differences between electrodes on the scalp, typically 10 to 100 microvolts, roughly a thousand times smaller than the voltage of an AA battery's millivolt range. The interesting part is where those voltages come from.

It's mostly not action potentials

The intuitive guess is that EEG picks up neurons "firing," but action potentials contribute surprisingly little. They last only about a millisecond, so thousands of them rarely line up closely enough in time to sum. And the current flows in and out of the axon in a way that largely cancels out at a distance.

The main source: synaptic currents in pyramidal neurons

The dominant contributors are postsynaptic potentials, the slower voltage changes (tens of milliseconds) that occur in a neuron's dendrites when it receives input. The key players are pyramidal neurons in the cortex, for two reasons.

First, their geometry. Pyramidal cells have a long apical dendrite running perpendicular to the cortical surface, and they're packed side by side in parallel, like trees in a planted forest.

Second, the current forms a dipole. When a synapse activates on the dendrite, positive ions flow into the cell at that point (a current "sink"). Current then flows along the inside of the dendrite and exits elsewhere along the cell (a "source"). This creates a tiny electrical dipole, with one end negative and the other positive relative to the surrounding tissue.

A single dipole is far too weak to detect. But when tens of thousands to millions of parallel pyramidal cells receive synchronized input, their dipoles add together. Estimates suggest several square centimeters of synchronously active cortex are needed for a clear scalp signal. This is why EEG is fundamentally a measure of synchrony: a million neurons active at random produce almost nothing, while a smaller population in lockstep produces a clear signal.

Strictly speaking, what flows is ions (sodium, potassium, chloride, calcium) moving through water, not electrons in a wire. The voltage at the scalp arises because those ionic currents spread through the conductive tissue around them.

Getting from cortex to scalp

The currents spread through the brain, cerebrospinal fluid, skull, and scalp, a process called volume conduction. The skull is a poor conductor, so it weakens and smears the signal. Each electrode therefore picks up a blurred mix from a wide area, which is why EEG has excellent timing (milliseconds) but poor spatial resolution (centimeters).

Some consequences

• Deep structures are nearly invisible. The hippocampus, thalamus, and brainstem are far away and not organized in the right parallel geometry, so their activity barely reaches the scalp directly. The thalamus does help drive many cortical rhythms, including sleep spindles and alpha, but what EEG records is the cortical response.
• Orientation matters. Cortex on the crest of a fold (gyrus) produces dipoles pointing toward the scalp, which EEG sees well. Cortex in the walls of folds (sulci) produces dipoles pointing sideways. MEG, which measures magnetic fields, is more sensitive to these, so the two techniques complement each other.
• Not all voltage on the scalp comes from the brain. Muscles in the jaw and forehead, eye movements (the eyeball itself is a dipole), and the heart all produce signals, often larger than the brain's. Removing these artifacts is a big part of EEG analysis.

So EEG is essentially listening to the summed synaptic input arriving at large, synchronized populations of cortical pyramidal neurons, filtered through several layers of tissue and bone.