Recording the P300 with EEG

Electroencephalography, EEG, records the electrical activity of the brain through electrodes on the scalp. The signal comes mainly from large groups of cortical neurons whose synaptic currents flow in parallel; by the time it reaches the scalp it is a few tens of microvolts (millionths of a volt), mixed with activity from the eyes, muscles and the mains supply (50 Hz in India, Europe and most of Asia and Africa; 60 Hz in most of the Americas). Good contact between each electrode and the skin, measured as impedance, is the first defence against this noise.

The response of the brain to a single sound is much smaller than the ongoing EEG. Researchers therefore present the same kind of event many times, cut the recording into short epochs aligned to each event, and average them. Activity that is time-locked to the event survives averaging; the rest cancels. The result is an event-related potential, or ERP, a series of positive and negative waves at characteristic times after the event.

In the auditory oddball task, frequent standard tones are mixed with rare target tones that the listener has to count. The targets evoke a large positive wave over the back of the head about 300 milliseconds or more after the sound, the P300 (Sutton et al., 1965). It is larger when the target is rarer and when the listener attends to it, and it is largest at parietal sites such as Pz (Polich, 2007).

The session follows one simulated volunteer from consent to a measured P300, and then a simulated group. Each step is a step of a real EEG study, and each choice changes the data as it would in practice. The core of the model, the ERP waves, background EEG and eye blinks, follows the EEG simulator; the effects of impedance, a gel bridge, a shifted cap, the Cz reference, more blinking without a blink instruction and a smaller P300 in passive listening are added as approximate illustrations. The simulated P300 is also larger and cleaner than in most real studies, so real experiments usually need more participants than the simulation suggests.

Learning objectives

• Describe what informed consent and ethical approval require before an EEG session, and how EEG data are pseudonymised and stored securely
• Locate Fz, Cz and Pz by the 10–20 system from head measurements, and explain the role of the reference and the ground
• Explain why electrode impedance matters and recognise mains interference, eye blinks, flat and bridged channels in raw EEG
• Design an oddball task and write an analysis plan (electrode, time window, rejection rule) before recording
• Epoch, baseline-correct, filter, reject artefacts and average EEG, and explain why noise falls with the square root of the number of trials
• Measure the P300 as a mean amplitude, test it across participants, and interpret the confidence interval and statistical power

Ethics and safety

Research with people requires approval by an ethics committee (in India, an Institutional Ethics Committee following the ICMR National Ethical Guidelines of 2017; elsewhere, a research ethics committee or, in the United States, an institutional review board) before recruitment begins, in line with the Declaration of Helsinki, and written informed consent from each participant after they have read an information sheet and had the chance to ask questions. Taking part is voluntary: participants may withdraw at any time without giving a reason and without any penalty, and the information sheet states whether, and until when, data already collected can be withdrawn (for example, until the data are anonymised or analysed). Exclusion criteria (for example hearing problems, neurological conditions, or medication that acts on the nervous system) are screened, and only the personal data the protocol needs are collected. Recordings are stored under a participant code, with the consent form, the only link between name and code, kept separately. Pseudonymised data of this kind still count as personal data: they are stored securely with access limited to the research team, kept only for the period stated in the information sheet and shared only in the way participants agreed to, in line with data protection law (in India, the Digital Personal Data Protection Act 2023; in the European Union, the General Data Protection Regulation). EEG is non-invasive and low-risk, but the gel and gentle skin abrasion can irritate sensitive skin, electrodes are cleaned and disinfected between participants, and the equipment must meet electrical safety standards for use with people.

Key terms

• EEG: Electroencephalography: recording the brain’s electrical activity with electrodes on the scalp.
• Electrode impedance: The opposition of the contact between an electrode and the skin to a small alternating current, in kilohms (kΩ). Low impedance means good contact; high impedance lets in mains interference and drift.
• 10–20 system: The international scheme for placing electrodes at 10% and 20% steps of distances measured on each head, from nasion (top of the nose) to inion (back of the skull) and from ear to ear.
• Reference: The electrode against which every channel is measured; each EEG channel is a voltage difference between its electrode and the reference. Here the linked mastoids, behind the ears.
• Oddball task: A sequence of frequent standard stimuli with rare target stimuli in random order, which the participant counts or responds to.
• Event marker: A code written into the recording at the moment each stimulus starts, so that the analysis knows when each event happened.
• Epoch: A short segment of the recording cut out around one event, here from 200 ms before to 800 ms after each tone.
• Baseline correction: Subtracting from each epoch its mean voltage in the period before the event, so that all epochs start from the same zero.
• Low-pass filter: A filter that keeps slow activity and removes faster frequencies; a 30 Hz low-pass keeps the ERP and removes most mains interference (50 Hz in India, Europe and most of Asia and Africa; 60 Hz in most of the Americas).
• Artefact: A signal that does not come from the brain, such as an eye blink, a muscle twitch, movement or mains interference.
• Event-related potential (ERP): The average voltage waveform time-locked to an event, obtained by averaging many epochs.
• P300: A positive ERP wave about 300 ms or more after a rare, task-relevant stimulus, largest over parietal sites.
• Mean amplitude: The average voltage of a waveform in a fixed time window, the usual way to measure an ERP component.
• Statistical power: The probability that a study detects an effect of a given size if it is real; it rises with the number of participants and, through cleaner averages, with the number of trials.